FIXEd aNd mOBIlE BROadBaNd

BROADBAND IN BRAZIL

Peter Knight Flavio Feferman Nathalia Foditsch

[eds.]

BROADBAND IN

braZil past, present and future

SÃO PAULO, 2016

Broadband in Brazil: past, present and future This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA. Original title: Banda larga no Brasil Copyright © 2016 by Peter Knight, Flavio Feferman, Nathalia Foditsch Copyright © 2016 by Novo Século Editora Ltda. editorial coordination

Vitor Donofrio editorial

Giovanna Petrólio João Paulo Putini Nair Ferraz Rebeca Lacerda

rights manager

Renata de Mello do Vale rights assistant

Acácio Alves

translation

cover art

Mike Gannan

Giovanna Foditsch

diagramming

João Paulo Putini

Cataloging-in-Publication Data (cip) Broadband in Brazil: past, present and future [e-book] / Peter Knight, Flavio Feferman, Nathalia Foditsch, orgs.; translation: Mike Gannan. Barueri, SP: Figurati, 2016. Original title: Banda larga no Brasil: passado, presente e futuro e-ISBN: 978-85-67871-78-3 1. Broadband communication systems 2. Computer networks 3. Internet 4. Telecomunication I. Knight, Peter II. Feferman, Flavio III. Foditsch, Nathalia IV. Gannan, Mike 16‑1290

cdd‑621.382181

Índices para catálogo sistemático: 1. Broadband communication systems

novo século editora ltda. Alameda Araguaia, 2190 – Bloco A – 11o andar – Conjunto 1111 cep 06455‑000 – Alphaville Industrial, Barueri – sp – Brasil Tel.: (11) 3699‑7107 | Fax: (11) 3699‑7323 www.novoseculo.com.br | [email protected]

CONTENTS Acknowledgments Preface Marília Maciel e Luiz Moncau – CTS/FGV Direito Rio foreword Peter Knight – Fernand Braudel Institute Flavio Feferman – UC-Berkeley Nathalia Foditsch – American University overview 1. The role of broadband in the development of Brazil Peter Knight – Fernand Braudel Institute 2. Fixed and mobile broadband: what is the future of the market? Eduardo Tude – Teleco Policies and regulation 3. National Broadband Program and Broadband for All: civil society’s perspective Veridiana Alimonti – Intervozes 4. General Plan of Competition Goals Abraão Balbino e Silva – Anatel 5. From scarcity to abundance: the debate on the efficient use of the electromagnetic spectrum Nathalia Foditsch – American University and Luca Belli – FGV 6. Access to Information and Communication Technologies in Brazil: a perspective of the demand for broadband Alexandre Barbosa, Alisson Bittencourt, Fábio Senne and Winston Oyadomari – Cetic.br 7. Policy and regulation: achievements and challenges Flávia Lefèvre – Proteste

8. Taxation of broadband services Peter Knight – Instituto Fernand Braudel Institutions 9. The rnp network and new partnerships Eduardo Grizendi and Michael Stanton – RNP 10. The role of NIC.br in the expansion of broadband infrastructure and in service quality improvement Demi Getschko – NIC.br 11. The contribution of small and medium operators and the role of Abrint Basilio R. Perez and Breno Vale – Abrint 12. The contribution of the large and medium operators João Moura – TelComp CASE STUDIES 13. Public-private partnerships for expanding broadband access: lessons from the Cinturão Digital do Ceará network in Brazil Fernando Carvalho – UFC, Flavio Feferman –UC-Berkeley, Peter Knight – Instituto Fernand Braudel and Glenn Woroch – UC-Berkeley 14. Underwater cables in Brazil and the world Eduardo Grizendi – RNP, Michael Stanton – RNP and Decílio Sales – Brazilian Army 15. Fiber in the favela: our experience in Rocinha Asafe Coimbra and Samuel Silva – Net Rocinha 16. Conclusions Flavio Feferman – UC-Berkeley about the authors

PRAISE FOR BROADBAND IN BRAZIL “The history of the Internet is different in each country. The trajectory of broadband in Brazil presented in this book made me admire the tactics and strategies that were successful in this country. We can all learn from this in-depth analysis.”

VINT CERF, VICE PRESIDENT AND CHIEF INTERNET EVANGELIST, GOOGLE “This book fills a great gap, documenting the history, challenges and achievements of the Internet in Brazil. The authors, many of whom have been active participants in its construction from the beginning, publish their contributions at a good moment, since the topic of the ‘Internet’ continues to be a current one, often polemical and provocative. Nothing better than to register this history to address current issues and confront future challenges. Further, I consider the book to be required reading for all who use the Internet in their daily lives.”

HARTMUT GLASER, EXECUTIVE SECRETARY, MANAGING COMMITTEE OF THE INTERNET IN BRAZIL (CGI.BR) “Enhancing understanding and awareness of society regarding the role of communication policies, institutions, and Brazilian companies has become essential for us to overcome the difficulties and demands we still face to bring quality broadband to all people. Examples and carefully gathered testimonies in this book demonstrate this can be done, and allow us to envision the role of the state and civil to realize enjoyment of this right for all Brazilians.”

NELSON SIMÕES, GENERAL DIRECTOR, NATIONAL RESEARCH AND EDUCATION NETWORK (RNP) “The spread of permanent and high-quality connectivity, compatible with the demands of users (today ‘broadband’ means access to a broad range of services involving massive use of multimedia content), is a challenge for all countries, including the most developed. A strategic plan that merits the name needs to embrace all the topics addressed in this book – an indispensable reference for planners, public entities, students, social and political leaders who seek quality (and hopefully ‘future-proof’) and universal access to the Internet.”

CARLOS ALBERTO AFONSO, REPRESENTATIVE OF THE THIRD SECTOR, BRAZILIAN INTERNET STEERING COMMITTEE (CGI.BR)

“This book is a major reminder of the importance of broadband as an unquestionable development tool for an emerging power like Brazil. We welcome this excellent contribution. The authors offer a thorough overview of the situation of broadband in Brazil today and the steps necessary to make it successful.”

Lourdes Casanova, Senior Lecturer and Academic Director, Emerging Markets Institute, S.C. Johnson School of Management, Cornell University “The quality of public policies depends on the availability of evidence and solid analyses of the various alternatives in debate. This book is a most valuable contribution to the improvement of government policies promoting the development and universalization of broadband in Brazil. The wide range of topics analyzed and the variety of perspectives of the authors make this book required reading for understanding the present situation and possible paths forward for the Internet in this country.”

Hernan Galperin, Research Associate Professor at the Annenberg School for Communication, University of Southern California “Brazil has been a regional and global leader in ICTs and broadband policy, strategy and tactics. This book, edited and authored by thought leaders in the field, comes at a very important and historical moment, when countries are willing to find ways to connect the ones that are not yet connected to the Internet. Countries committed to the United Nations, the ITU’s ‘Connect 2020 Agenda’ – will make efforts to bring the next 1.5 billion people online by 2020 and to use ICTs as a way to achieve and support the Sustainable Development Goals (SDGs). The documentation and analysis brought by this book will surely support these efforts.”

Carolina Rossini, Vice President for International Policy and Strategy at Public Knowledge and World Economic Forum Young Global Leader

acknowledgments

T

his book is the collective work of 23 authors. They are the first people we would like to thank for the care and professionalism shown during the nine months elapsed from the first drafts to the delivery of the Portuguese language manuscript to Editora Novo Século. They all worked on a voluntary basis, driven only by their interest in the subject and their will to contribute to the debate on the expansion of broadband and its role in the development of Brazil. We thank Michael Stanton, from the National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP), and José Roberto Afonso, for proofreading Chapters 1 and 8, respectively. We would also like to thank Zaida Gusmão Knight, for the linguistic and grammar revision of the Portuguese version of these chapters, and Deborah R. Zimmerman, for her contribution with ideas and text revision. We would also like to express our thanks to the team of Editora Novo Século – Cleber Vasconcelos, Vitor Donofrio and João Paulo Putini – for their receptiveness and professionalism, which were extremely valuable in the pro‑ duction of this book. They understood the relevance of our task for Brazil and were able to produce the book in a record time. Global Nexus did the initial translation into English, financed by Google. We are also thankful to the coordinators of the Center of Technology and Society (Centro de Tecnologia e Sociedade – CTS) at the Fundação Getulio Vargas (FGV-RIO) Law School: Marília Maciel and Luiz Fernando Moncau, who supported the idea of writ‑ ing this book from the very beginning. The support of Sacha Leite and Thaís Mosque, both responsible for publications of the Law School of FGV-RIO, was also vital for this project. The book cover was developed by designer and illustrator Giovanna Foditsch, whom we thank for her creativity and the job she did. Jaime Rui de Sousa Júnior and Vinicius Araújo Oliveira dos Santos, from the RNP’s Coordination Unit for Design, Corporate Communication Management, con‑ tributed with the high-resolution maps presented in Chapters 9 and 14. Finally, we would like to thank our families for the support and affection they gave us while we were working on this book.

PREFACE

A

ccording to the World Bank’s World Development Report 2016, Brazil ranks fifth in number of Internet users in the world – after China, the United States, India and Japan, in this order. Although in absolute numbers this is a significant accomplishment, there is still a considerable portion of the population lacking broadband access. Moreover, other issues –such as the quality of connections – undermine these people’s possibility to fully enjoy the benefits of the information society. Nevertheless, despite these enduring disparities, broadband is increasingly present in our lives. The Internet permeates many of our personal relationships, our political participa‑ tion, and other daily interactions and activities. When it comes to companies, the broadband Internet enhances productivity, competitiveness, and innova‑ tion, enabling the creation of new products and services. In government, the broadband Internet expedites information manage‑ ment, improving public administration and enabling a series of e-government services. And for each one of us, broadband expands our access to education, healthcare, entertainment, communication, and participation in society. Full achievement of this transforming capacity depends on well-balanced deci‑ sions regarding public policies that promote rights and encourage innovation. In developing countries, some of the key policies currently under discussion are related to the provision of broadband. The analysis of public policies, regulation, practical cases, and market characteristics is critical for us to make progress and enhance the debate on broadband expansion. Broadband in Brazil – past, present and future is an im‑ portant contribution to this analysis, as it provides different perspectives on the subject. It investigates Brazilian examples, like that of a visionary access provider in Rocinha, a large squatter community in Rio de Janeiro, regional experiences, like Ceará’s Digital Beltway, and underwater fiber optic networks in the Amazon. But it also analyzes problems of national importance, like the progress achieved and difficulties of the National Broadband Program, chal‑ lenges of the General Plan of Competition Goals, and disputes concerning net neutrality.

12

BROADBAND IN BRAZIL

Guided by authors who are leaders in the debate on broadband in Brazil, the reader will gain not only a broad overview of the subject, but also a deep, technical, and revealing approach to the role of some fundamental institu‑ tions for the development of the counry’s broadband market. This book is a timely reminder that despite all the achievements to date, we still have a long way to go in Brazil. The Center of Technology and Society (Centro de Tecnologia e Sociedade – CTS) of the Fundação Getulio Vargas’s Rio de Janeiro Law School, that we are proud to coordinate, has been doing research to enrich the debate in this area. We firmly believe that the debate about the Internet in Brazil and the world should be carried out from multiple perspectives, with the participation of governments, civil society, academia, technical community, and companies. Diversity is the hallmark of this book. We believe that only from a plural point of view can we develop the web while preserving the rights and guaran‑ tees of our citizens. Access to broadband is essential for development of our society. The policies aimed at assuring it are challenging and have no easy solutions. This book presents an essential contribution to this debate. Luiz Moncau and Marília Maciel Co-managers Center of Technology and Society of the Fundação Getulio Vargas Law School in Rio de Janeiro (CTS/FGV Direito Rio)

Foreword

T

he lights of the router are blinking. It is the Wi-Fi communicat‑ ing with our various devices that depend on its signal. We send e-mails, access our social networks, check the weather forecast and are bombarded by ads, news, and information every single day. Soon, not only will our cell phones and computers be connected to the Internet, but also our air-conditioning systems, our refrigerators, and even our cars. Where does this come from? The Internet seems like magic to most of us, but it depends on infrastructures and institutional arrangements, pub‑ lic and private investment, policies, and regulation. Without this intricate ecosystem, the incredible innovations enabled by the broadband Internet would not be possible. The Internet has been a reality in Brazil for over twenty years. Its growth has been driven by the government, large companies, small and medium-sized providers, as well as by increased demand. The accomplishments so far are evident, but a large portion of the population is still not included in the digi‑ tal world, lacking access to the Internet. Furthermore, the cost of broadband remains high and its quality must be improved so that we can make use of all the benefits it can bring, benefits related to healthcare, education, and even political rights. The Brazilian Internet must, therefore, be fast, reliable, and affordable to spur economic growth as well as to promote citizens’ social and political rights. To achieve these objectives, we need to discuss some develop‑ ment strategies for broadband in Brazil. Large economies of scale in Internet infrastructure are an important fac‑ tor to be considered in elaborating broadband development strategies. These economies of scale result from high fixed costs, but low marginal costs to provide increasing amounts of goods or services. Barriers to entry, technolog‑ ical convergence, and the vertical integration of operators in the markets for telecommunications services are some of the reasons why these markets are subject to the power of only a few companies. These dynamics lead us to a debate on industry regulation, public policies, and the role of the government in universalizing broadband Internet access. Lowering the costs of broadband services and promoting the development of well- regulated and competitive

14

P. KNIGHT • F. FEFERMAN • N. FODITSCH

markets are essential for us to achieve universal access. Additionally, to really achieve the economic, social, and political objectives of broadband, we must address other matters, such as the development of infrastructure, a system of content and services, as well as encouragement to their use. This book was written by specialists in several areas related to broadband who agreed to share their experience with public policies and regulation in Brazil, including experiences of entrepreneurship and public-private partner‑ ships in the country. This debate is fundamental for the technological, eco‑ nomic, and social progress of Brazil.

EXECUTIVE SUMMARY This book is divided into four parts: (i) overview; (ii) policies and regula‑ tion; (iii) institutions; and (iv) cases studies. The last chapter presents some conclusions based on the contributions of all the authors.

(i) Overview The first section presents an overview in two chapters. Chapter 1, by Peter Knight, reviews the literature on the impact of broadband on development and analyzes the available data on the status of broadband in Brazil as well as federal public policies in the country. In Chapter 2, Eduardo Tude presents definitions and basic concepts to be used throughout the book, in addition to an analysis of the present and the future of fixed and mobile broadband markets in Brazil.

(ii) Policies and regulation The second section starts with Chapter 3, by Veridiana Alimonti, pre‑ senting a civil society perspective of the National Broadband Program and new public policies that are being elaborated. In Chapter 4, Abraão Balbino e Silva explains the theory behind, the objectives of, and implementation of the General Plan of Competition Goals, aimed at increasing competitiveness and lowering prices of telecom services. In Chapter 5, Nathalia Foditsch and Luca Belli describe the importance of electromagnetic spectrum manage‑ ment for wireless broadband development. In Chapter 6, Alexandre Barbosa, Fábio Senne, Alisson Bittencourt and Winston Oyadomari use data from from the Regional Center of Studies for the Development of the Society of

presentation

15

the Information (Cetic.br) to analyze the differences in access and use of the Internet in Brazil and suggest relevant aspects for the creation of new digital inclusion policies in the country. Chapter 7, by Flavia Lefèvre, presents comments on the achievements and challenges of public policies and regulation, with emphasis on the Brazilian Civil Rights Framework for the Internet (Marco Civil da Internet) and the de‑ bate on net neutrality. In Chapter 8, Peter Knight argues that the telecommunications indus‑ try is subject to an extremely heavy tax burden and that most of the tributes collected by the three federal funds created to support the sector are used for other purposes, not contributing to its development.

(iii) Institutions The third section of the book analyzes the contribution of important institutions that participate in the development of broadband in Brazil. In Chapter 9, Michael Stanton and Eduardo Grizendi examine the role played by the National Research and Education Network (RNP) ever since its foun‑ dation, as well as the development of its fiber optic backbone, the programs of metropolitan networks – Community Networks of Education and Research (Redecomep) – and the extending these networks to reach higher education and research institutions in Brazil’s interior. In Chapter 10, Demi Getschko examines the role of the Brazilian Network Information Center (NIC.br), the executive arm of the Brazilian Internet Steering Committee (CGI.br), in expanding broadband infrastructure and improving service quality. Getschko explains the role of the various entities associated to the NIC.br and the challenges of the future. One of the distinguishing characteristics of the Internet in Brazil is the presence of over 3,000 small and medium-sized Internet providers. In Chapter 11, Basílio Perez and Breno Vale show the origin and the evolution of these providers, their contribution to the digital inclusion of the population, the challenges they face, and the role of the Brazilian Association of Internet and Telecommunications Providers (Abrint) in the defense of these companies. The Brazilian Association of Competitive Telecommunication Service Providers (TelComp) is another important sector organization that congre‑ gates several telecom operators. Its contribution is addressed in Chapter 12, written by João Moura. The chapter includes an historical overview of com‑ petition in the telecommunications industry and highlights the importance

16

P. KNIGHT • F. FEFERMAN • N. FODITSCH

of regulation to avoid abuses by companies with Significant Market Power (SMP), with the objective of lowering prices and encouraging the develop‑ ment of new services.

(iv) Case studies In the fourth section of the book, we present case studies of innovative initiatives that merit attention on the part of the civil society, the private sec‑ tor, regulators, and public policy makers. In Chapter 13, Fernando Carvalho, Flavio Feferman, Peter Knight and Glenn Woroch analyze the lessons we can learn from the experience of Ceará’s Digital Beltway (CDC), a statewide fi‑ ber optic network with extensions via radio links. The CDC is an initiative of Ceará’s Information Technology Company (Etice) and the RNP, in collabora‑ tion with Ceará’s Eletric Power Company (Coelce). The authors explain that Etice has developed a new model of public-private partnership (PPP) that was able to attract private companies to use and expand this network. The Connected Amazon program – an initiative involving the Brazilian Army, the RNP and other partners – uses underwater cables to bring broad‑ band access to remote areas in the Amazon. This experience is analyzed by Eduardo Grizendi, Michael Stanton and Decílio Sales in Chapter 14. The pro‑ gram intends to deploy 8,000 kilometers of fiber optic cables in the rivers of the Amazon. The technology developed and tested by the program can be used in other regions with large bodies of water and difficult overland access. In Chapter 15, Samuel Silva and Asafe Coimbra present the experience of Net Rocinha, which enabled the digital inclusion of thousands of inhabitants of Rocinha, one of the biggest favelas in Brazil. The authors tell a bit of their own history and how they created and formalized their company, which operates today under a license from the national telecommunications regulator, Anatel.

Conclusions Finally, in Chapter 16, Flavio Feferman shows how the ideas contained in the fifteen previous chapters can contribute to the development of the broad‑ band system in Brazil. Peter Knight – Fernand Braudel Institute of World Economics Flavio Feferman – Developing Markets Group and University of California, Berkeley Nathalia Foditsch – American University

overview

|1|

THE ROLE OF BROADBAND IN THE DEVELOPMENT OF BRAZIL Peter Knight [ Fernand Braudel Institute of World Economics ]

W

introduction

idespread high-speed, high-quality and low-cost broadband is critical to expedite economic, social and political development in Brazil. The country’s competitiveness within a globalized econo‑ my increasingly depends on knowledge and access to information. Broadband Internet is essential to the infrastructure of the 21st century. It leverages tech‑ nological advances in several areas, reducing costs and increasing the qual‑ ity of services such as education, health and public security, among others. Broadband can also strengthen the research and education system, promote innovation and foster economic development, contributing to a more just so‑ ciety and a more dynamic and competitive economy. Extending fiber optic networks to almost all municipalities and building high-capacity networks that reach the most remote remaining municipal‑ ities must be one of the main thrusts of a broader strategy for the use of Information and Communication Technologies (ICT) to promote the coun‑ try’s economic, social and political development. Note that even when the user accesses the Internet via wireless technol‑ ogies (fixed or mobile), a high capacity backhaul is necessary, either via ra‑ dio links or fiber optics, to reach the networks that are the backbones of the Internet. How to accomplish these goals is part of a national debate that has two main elements. The first is the development of the Banda Larga para Todos (Broadband for All – BLPT) program, proposed during the reelection cam‑ paign of President Dilma Rousseff, in 2014. The goals and the structure of the program were still unclear and under discussion in February, 2016. The

20

peter knight

second is about the so called “new telecommunications model”, subject of a public consultation by the Ministry of Communications (Ministério das Comunicações – MC). Despite significant advances, broadband in Brazil remains slow, expen‑ sive and often of lower quality when compared to other countries. In Brazil there are also major inequalities of access, quality and cost of broadband. Developing a broadband strategy for Brazil is essential, as well as is giv‑ ing it political priority to mobilize the necessary resources to remedy these shortcomings. This chapter reviews recent research on the contribution of ICT to devel‑ opment, focusing on broadband Internet. It also summarizes the current situ‑ ation of broadband in Brazil in comparison with other countries and analyzes some of the main issues in the debate on broadband policies.

INTERNET AND ECONOMIC, SOCIAL AND POLITICAL DEVELOPMENT The power of ICT affects almost every aspect of today’s economy and society. The World Development Report 2016 of the World Bank, issued in January, 2016 (WORLD BANK, 2016), is the most comprehensive and up‑ dated study on the digital revolution’s impact on digital transformation strat‑ egies adopted by different countries. It also addresses the contribution of the Internet to accelerate economic, social, and political development. Its main findings are (World Bank, 2016, p. 4): Access to the Internet is critical, but not sufficient. The digital econ‑ omy also requires a strong analog foundation, consisting of regu‑ lations that create a vibrant business climate and let firms leverage digital technologies to compete and innovate; skills that allow work‑ ers, entrepreneurs, and public servants to seize opportunities in the digital world; and accountable institutions that use the Internet to empower citizens. […] The triple complements – a favorable business climate, strong hu‑ man capital, and good governance – will sound familiar – and they should because they are the foundation of economic development. But digital technologies add two important dimensions. First, they raise the opportunity cost of not undertaking the necessary reforms. They amplify the impact of good (and bad) policies, so any failure

the role of broadband in the development of brazil

21

to reform means falling farther behind those who do reform. With digital technologies, the stakes have risen for developing countries, which have more to gain than high-income countries, but also more to lose. Second, while digital technologies are no shortcut to devel‑ opment, they can be an enabler and perhaps an accelerator by rais‑ ing the quality of the complements.

But a necessary condition to materialize the digital dividend is the ex‑ istence of a broadband infrastructure, which, in turn, requires massive in‑ vestments. Another World Bank study (QIANG; ROSSOTTO, 2009) shows a positive relationship between economic growth and the pervasiveness of telecommunications, in which broadband is the most impacting.

FIGURE 1.1. How the investment in broadband and information technology can impact economic growth 1,6 1.6

1.38

1,4 1.4 1.12

1.2 1,2 1

0.6 0,6

0.81

0.73

0.8 0,8

1.21

0.77

0.6 0.43

0,4 0.4 0,2 0.2

0 Fixed

Mobile

High-income economies

Internet

Broadband

Low and medium-income economies

Note: the vertical axis is the increase in the economic growth rate for each 10% increase in penetration Source: Qiang and Rossotto (2009, p. 45)

Similar results were presented in a report published by the International Telecommunications Union (ITU) (Figure 1.2), which shows what appears to be a direct relationship between broadband penetration and the magnitude of the impact on economic growth.

22

peter knight

Contribution to GDP growth for each additional 10% penetration 30%

FIGURE 1.2. The relationship between broadband penetration and the importance of its contribution to GDP growth 30%

Germany – High penetration

25%

Germany – Low penetration

Arab States

20% 15%

Latin America

10%

OECD – Average penetration Chile

Brazil

5%

OECD – High penetration

OECD – Low penetration

0% 0%

5%

10%

15%

20%

25%

30%

35%

40%

Broadband penetration Source: Katz (2012, p. 93).

In short, the development of the digital economy favors the develop‑ ment of analog complements, and vice versa. It is a feedback process, a reciprocal interaction, a positive spiral that produces what the World Bank calls a “digital dividend”. Social and political development is also favored by digital inclusion, enabling greater participation of citizens and civil society organizations in political activities, both via social networks or via public consultations conducted online. This participation can also enhance economic development. The reciprocal relationship between digital development and its analog complements is summarized in Figure 1.3. To fulfill the potential of the dig‑ ital revolution, that is, to receive the digital dividend, we must have a digital transformation strategy that embraces both the expansion of connectivity and the development of analog complements.


23

FIGURE 1.3. Why digital dividends do not spread quickly – and what can be done about it

Source: World Bank Group (2016, p. 4).

Without a ubiquitous broadband infrastructure, social and economic in‑ equality can only grow. Why? The lower classes are digitally excluded – either because they are poor, because they lack digital literacy or because they live in remote areas where broadband infrastructure is not available. Therefore, pro‑ moting digital inclusion should be a central goal of a digital transformation strategy. But broadband is only a tool to achieve broader development goals. It is the ICT applications that take the benefits to citizens, businesses, govern‑ ment institutions, and third sector entities. Therefore, a strategy to promote and leverage investment in broadband requires a holistic approach that also includes the development of institu‑ tions and policies to promote digital inclusion, education and training, in addition to creating a business environment that fosters innovation. Such a strategy is essential for us to realize substantial synergies and economies of scale inherent to the use of ICT (HANNA; SUMMER, 2015). For exam‑ ple, advances in education and research lead to better health standards in the population; greater citizen participation can result in better public pol‑ icies; and large data centers can serve several government agencies with more security, lower prices and lower cost per user (whether civil servants, businesses or citizens) than smaller servers distributed across the same government entities. Developing and implementing such a strategy is not easy, because gov‑ ernments are organized in ministries and departments, each defending its

24

peter knight

territory and resisting attempts of coordination. Countries that have adopted this strategy (South Korea, Singapore and Finland are good examples) had political leaders with a vision, political will and determination to implement it and had or helped create a national consensus to support it (HANNA; KNIGHT, 2011, 2012). The only government attempt to develop such a strategy in Brazil was car‑ ried out under the coordination of the Ministry of Science and Technology. It was published in the year 2000 as Information Society in Brazil: Green Book (Sociedade da Informação no Brasil: Livro Verde) TAKAHASHI, 2000). It has never been implemented. Concerned with the absence of an official program, 70 authors from the public and private sectors, civil society and academia, joined forces in the e-Brasil Project. In 2006 and 2007, they published two books uniting their contributions, with guidelines and actions on what they called the e-Brasil Program. The aim was to trigger a national debate on digi‑ tal transformation in the country (KNIGHT; FERNANDES, 2006; KNIGHT; FERNANDES; CUNHA, 2007). Nevertheless, Brazil still lacks an official strategy for digital transformation.

CURRENT STATUS OF BROADBAND IN BRAZIL Table 1.1 presents some data on Internet penetration and the capacity of access connections for the years 2006/2014.1 The statistics show a progres‑ sive improvement in Internet access. However, in 2014 only 50% of house‑ holds had access to any type of Internet connection and only 10% had a con‑ nection above 2 Mbps. Among individuals over 10 years old, 44% accessed the Internet at least once a day in 2014. Of all Internet users, 76% accessed it via smartphones, almost as many as those who accessed it via computers, laptops or tablets (80%).

1 By capacity of access connection we mean the maximum rate of transmission of information available between the end user and his/her connection provider. It depends on the technology used for this connection and may be asymmetrical, with greater capacity to receive (download) than to transmit (upload).


25

TABLE 1.1. Internet and broadband penetration, Brazil, 2006/2014 Statistics/Year

2006

2008

2010

2012

2013

2014

% private households with Internet access

15

20

27

40

43

50

% private households with Internet access with landline (fixed) broadband connectiona

6

10

18

27

28

34

% private households with Internet access with mobile broadband connection (3G modem)

n.a.

n.a.

3

8

9

13

% private households with Internet access with connection between 2 and 8 Mbpsb

n.a.

1

3

3

6

8

% private households with Internet access over 8 Mbps

n.a.

n.a.

1

6

7

12

% of individuals over 10 years old who accessed the Internet in the last 3 months before any local search (considered Internet users in the research)

28

38

41

49

51

55

% of individual Internet users over 10 years old who accessed the Internet every day from any location

13

21

25

34

36

44

% of individuals over 10 years old who accessed the Internet anywhere in the three months prior to the survey via computer, laptop or tablet

n.a.

n.a.

n.a.

n.a.

n.a.

80

% of individuals over 10 years old who accessed the Internet in the three months prior to the survey via cell phone, from any location

n.a.

n.a.

n.a.

n.a.

n.a.

76

Source: CGI.br (2007/2015). Notes: n.a. = not available a. Broadband is defined as higher than 128 Kbps for Digital Subscriber Line (DSL), cable, radio or satellite. b. Transmission rates given only to households where the person who answered the research was aware of the speed.

More detailed data for the year 2014 clearly show that all of these indicators are lower in rural areas, in poor or remote regions (North and Northeast) and for individuals with income or education levels below average (CGI.BR, 2015 tab. A4, A6, and C1, p. 324, 322 and 334). For example, 98% of the A Class households had Internet access, but only 14% of the households in the D and E Classes. And so on: 54% of urban households vs. 22% rural, 62% in the Southeast vs. 35% in the North, 83% of individuals aged 16-24 years old vs. 15% of those over 60, and

26

peter knight

97% of individuals with higher education vs. 4% of illiterate or with primary edu‑ cation had access to the Internet, according to the survey. The same differences in access in urban and rural areas, social classes, major regions, age and education have existed since the beginning of such research in 2006, but all these indicators have improved over time.2 In short, digital inclusion has advanced substantially, but there is still much to be done to reach the poorest regions and poorer or and/or less edu‑ cated individuals, older age groups and rural areas.

Comparisons: Brazil and selected countries Brazil ranks low, according to various Internet service indicators, com‑ pared both to rich countries as well as other developing countries. For com‑ parison, we chose the so-called “BRICS” (Brazil, Russia, India, China and South Africa), three Latin American countries with per capita income similar to Brazil’s (Argentina, Mexico and Uruguay), as well as South Korea, a newly industrialized country and a leader in connectivity.

Penetration Table 1.2 presents various indicators of Internet penetration in the select‑ ed countries. The first indicator is the penetration rate of fixed broadband per 100 inhabitants. The rate of 11.5% puts Brazil well below leader South Korea (38.8%) and also below most countries with similar per capita income level (Uruguay, Russia, Argentina and Mexico) and China. South Africa and India exhibit much lower penetration rates. In the case of India, this is because of the huge rural population with low income and low education levels. South Africa faces many of the same problems of fixed broadband as Brazil, includ‑ ing underinvestment in infrastructure, little competition in the provision of broadband services and the high cost of connectivity. Table 1.2 includes broader indexes of information technology. These in‑ dexes include: (a) the ICT Development Index (IDI) of the International Telecommunication Union (ITU); (b) the sub-index of the IDI, which measures network access (IDI-Access);3 and (c) the Network Readiness Index (NRI) from 2

For details, see Chapter 6 of this book.

3

The IDI is a composite index that combines 11 indicators of access, use and skills in a measure used to monitor and compare developments in ICT in different countries. The IDI Access is a composite of five of these 11 indicators: landline subscriptions per 100 inhabitants, mobile phone subscriptions per 100 inhabitants, Internet bandwidth (Mbps)


27

the World Economic Forum.4 As for these broader indicators, among the BRICs Brazil is second only to Russia. Compared to the three Latin American coun‑ tries, Brazil is ahead of Mexico, in the same relative level of Argentina, however, well behind Uruguay in the ranking of these respected international indexes.

TABLE 1.2. Comparison of the ranking of Brazil and selected countries in the proportion of the population with fixed broadband, IDI, IDI Access Component, 2015 and NRI, 2015 Fixed broadband (% of the population) 2014a

IDI 2015 Ranking

IDI – Access 2015 Ranking

NRI 2015

Brazil

11.5

60

71

84

Russia

Country

17.5

45

48

41

India

1.2

131

135

89

China

14.4

82

89

62

3.2

88

88

75

Argentina

14.7

52

64

91

Mexico

11.6

95

98

69

Uruguay

24.6

49

50

46

South Korea

38.8

1

9

12

South Africa

Sources: ITU (2015) for fixed broadband (p. 222-225), IDI (tab. 2.2, p. 44) and IDI access (tab. 2.3, p. 46); Dutta; Geiger; Lavin (2015 tab. 1, p. 8) for NRI. Note: a. Fixed broadband is defined as a connection faster than 256 Kbps for download via DSL, coaxial cable, FTTx (fiber to the curb, building or home) or other fixed line.

Data download rate The faster the rate, the better and more effective the use of the Internet. Using Akamai’s statistics on the download rate from their web servers,5 the per Internet user, percentage of households with a computer and percentage of households with Internet access. For details on the methodology, see ITU (2014, Annex 1, p. 221-225). 4

The NRI is a composite index that has four main categories (sub-indexes), 10 subcategories (pillars) and 53 indicators distributed across various pillars. The main categories are the environment (political and regulatory, and business and innovation), readiness (infrastructure, cost and skills), use (individuals, businesses and governments) and impact (economic and social). For details on the methodology, see Dutta; Geiger; Lavin (2015, ch. 1, p. 37).

5

The indicators produced by Akamai measure the download of web content transferred with the TCP transport protocol (without data loss). For various reasons (competition with

28

peter knight

average rate for Brazil in the third quarter of 2014 was 3.6 Mbps. This rate is 71% of the global average, and Brazil ranks below most countries in the com‑ parison chart, particularly South Korea (20.5 Mbps) and Russia (10.2 Mbps), but also below other countries in Latin America – Uruguay (5.9 Mbps), Mexico (5.5 Mbps), and Argentina (4.2 Mbps). Brazil is at the same level of South Africa and China (3.7 Mbps), and is only ahead of India (2.5 Mbps). The data indicates a low connection speed in Brazil, which jeopardizes the use of a whole new generation of services and applications. According Akamai’s data, 32% of fixed broadband connections in Brazil had download speeds above 4 Mbps, compared to 39% in Argentina, 64% in Mexico, 68% in Uruguay and 96% in South Korea. The percentage of users with download rates above 10 Mbps was only 2.2% in Brazil, above the level of China, but below all other countries (Table 1.3). Only 0.6% of connections in Brazil had a download speed above 15 Mbps. These statistics reinforce the low quality of broadband access in Brazil.

TABLE 1.3. Comparison of download rates of fixed broadband in Brazil and selected countries, third quarter of 2015a Download rate average (Mbps)

% above 4 Mbps

% above 10 Mbps

% above 15 Mbps

5.1

65

27

15

Brazil

3.6

32

2.2

0.6

Russia

10.2

87

38

15

India

2.5

14

2.3

0.8

China

3.7

33

1.6

0.3

South Africa

3.7

22

2.9

1.7

Argentina

4.2

39

3.1

0.5

Mexico

5.5

64

6.4

1.7

5.9

68

7.7

1.6

20.5

96

68

45

Country Global average

Uruguay South Korea Source: Akamai (2015).

Note: a. Average speed of IP addresses connecting to Akamai servers, excluding wireless networks and cloud hosting services that typically have extremely fast Internet connections.

third-party transfers, data retransmission lost in the transfer, among others), this is usually well below the capacity of end-user access connection (AKAMAI, 2015b).


29

Prices As for the monthly prices customers pay for Internet access, Brazil’s sit‑ uation is improving, but it still falls short of expectations. The ITU has an annual publication (ITU, 2015) that shows prices in 2014 for many countries. The ITU sets a maximum level of 5% of the Gross National Income per capita spent on fixed broadband services. Brazil fits within that limit, even though the price per Mbps is high.6 Considering fixed broadband, the monthly price indicator for an entry lev‑ el (basic) package of services as a percentage of the monthly Gross National Income per capita, an indicator of accessibility to the average consumer, Brazil has the most affordable fixed broadband among all selected countries, except Russia. However, when measured in USD/Mbps, Brazilian prices are the most expensive, except for China and Argentina. According to this criteri‑ on, prices in Brazil are almost five times higher than in Uruguay.

TABLE 1.4. Monthly price and broadband connection capacity comparison in Brazil and selected countries, 2014a Country Global average Brazil Russia

Rating (ranking) % of monthly Gross National Income (GNI) per capitab

Price as % of monthly GNI per capita

n.a.

20.80

45

1.30

Price in PPP$

Mbps, basic access basket

USD/ Mbps

30.40

52.30

n.a.

n.a.

12.66

16.62

1.0 12.66

Price in US$

11

0.68

7.82

17.94

15

0.52

India

108

5.28

6.90

24.04

1.5

4.60

China

90

3.58

19.53

31.92

1.0

19.53

74

2.46

15.20

31.94

2.0

7.60

n.a.

n.a.

41.93

n.a.

3.0

13.98

South Africa Argentinac Mexico

82

3.17

26.26

37.30

5.0

7.46

Uruguay

29

1.08

13.64

17.03

5.0

2.73

South Korea

47

1.32

28.49

32.80

50.0

0.57

* n.a. = not available Source: ITU (2015, graph. 4.5, p. 107, tab. 4.4, p. 109).

6 In terms of fixed broadband, the user typically contracts a service with limited access connection. This may also be asymmetrical (with the download capacity higher than the upload capacity). The price is usually estimated per Mbps.

30

peter knight

Notes: a. Prices and access connection capacities for a basic basket of broadband services (entry level), prices in USD PPP (purchasing power parity) calculated by the World Bank method (see ). For the detailed methodology of price collection, see ITU (2015, Annex 3, p. 210, 212 and 213). b. The lower the rating, the more affordable to the population. c. Country not rated because of lack of data on the Gross National Income per capita in the last five years.

Consulting firm Teleco did a survey in October, 2015, using criteria similar to those of ITU to compare broadband prices in 18 countries. Maybe because it used a higher exchange rate (3.97 BRL/USD) from late September 2015 – the average rate for 2014 was 2.35 BRL/USD – Brazilian prices in USD were rather low, 59% of the ITU study prices for 2014, lower than those of 15 of the other 17 countries (Figure 1.4). It turns out that the average exchange rate for the year 2014 was 59% of the rate from October 30th, 2015. Converting prices to Brazilian reais gives us almost the same price in both studies: 29.75 BRL in 2014 and 29.77 BRL in October, 2015.

FIGURE 1.4. Fixed Broadband Package Price in USD, October 2015 52.6 40.3 32.4 30.6

27.1 24.8 24.5 23.4 23.3 21.5 20.6 18.2

15.0

12.0

7.6

7.5

5.3

4.2

Source: Teleco (2015a, p. 6).

Table 1.5 compares prices for broadband in mobile prepaid phones in the same countries of the ITU survey conducted in 2014. The prices in USD cents/MB are the highest of all compared countries, except for Argentina. They are 4.6 times higher than in Uruguay and 12.6 times higher than in Russia.


31

TABLE 1.5. Comparison of prices and franchise of mobile broadband data, prepaid cell phone in Brazil and selected countries, 2014a

Country

Brasil

Rating (ranking) % of Gross National Income (GNI) per capitab

Price as % of GNI per capita

54

1.13

11.00

MB

US Cents/ MB

14.43

500

2.89

Price Price in US$ in PPP$

Russia

14

0.45

5.21

11.96

2,250

0.23

India

88

2.48

3.24

11.29

600

0.54

China

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

66

1.48

9.12

19.16

500

1.82

n.a.

n.a.

17.34

n.a.

500

3.47

South Africa Argentinac Mexico

95

2.72

22.49

32.96

1,024

2.20

Uruguay

31

0.68

8.60

10.75

768

1.12

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

South Korea

* n.a. = not available Source: ITU (2015, tab. 4.8, p. 137). Notes: a. Prices and transferred data limits for a basic package (entry level), prices in USD PPP (purchasing power parity) calculated by the World Bank method (see ). For the detailed methodology of price collection, see ITU (2015, Annex 3, p. 210, 213 and 214). b. The lower the rating, the more affordable to the population. c. Country not rated because of lack of data on the Gross National Income per capita in the last five years.

Teleco made a comparative survey of prices of prepaid mobile broadband in October 2015, with a methodology somewhat different from the ITU study. It covered 15 countries, including Brazil. It considered promotional prices as long as they did not establish retention periods, for recharges with an expira‑ tion period of about 30 days. The prices collected by ITU exclude promotional offers. The results show lower prices than the ITU research from 2014, proba‑ bly because of the higher exchange rate and the use of promotional offers pric‑ es. In this study, Brazil had lower prices than 11 of the 15 countries included in the survey (Figure 1.5). Converting Brazilian prices into dollars of 2014 and October 2015, we have 25.85 BRL and 23.82 BRL, respectively.

32

peter knight

FIGURE 1.5: Prepaid Mobile Broadband Package for mobile phones in USD, October 2015 25.0 21.1 13.5 12.6 11.2 10.7

9.3

9.1

8.4

8.4

6.7

6.0

5.6

5.3

3.9

Source: Teleco (2015b, p. 6).

Postpaid cell phone prices paid by Brazilians, collected by ITU in 2014, are much higher than in any other country, according to any of the four crite‑ ria: 37 times those of Russia and 8 times those of Uruguay, measured in USD cents/MB (Table 1.6).

TABLE 1.6. Comparison of prices and franchise of mobile broadband data, postpaid cell phone in Brazil and selected countries, 2014a Rating (ranking) % of Gross National Income (GNI) per capitab

Price as % of GNI per capita

Price in US$

Price in PPP$

Brazil

114

4.32

42.07

Russia

14

0.45

5.21

Country

MB

US Cents/ MB

55.22

500

8.41

11.96

2,250

0.23

India

97

2.51

3.28

11.41

600

0.55

China

44

0.89

4.88

7.98

500

0.98

69

1.48

9.12

19.16

500

1.82

South Africa Argentina

c

n.d.

n.d

17.34

n.d.

500

3.47

81

1.81

14.97

21.27

500

2.99

Mexico Uruguay

42

0.85

10.71

13.38

1,024

1.05

South Korea

18

0.48

10.45

12.03

500

2.09

* n.a. = not available Source: ITU (2015, tab. 4.7, p. 136).


33

Notes: a. Prices and transferred data limits for a basic package (entry level), prices in USD PPP (purchasing power parity) calculated by the World Bank method (see ). For the detailed methodology of price collection, see ITU (2015, Annex 3, p. 210, 213 and 214). b. The lower the rating, the more affordable to the population. c. Country not rated because of lack of data on the Gross National Income per capita in the last five years.

Just like in the case of fixed broadband and prepaid mobile, Teleco’s re‑ search about postpaid mobile broadband prices conducted in October 2015 showed that prices for Brazil in USD are lower than those in the ITU survey for 2014, with Brazil having lower prices than 9 of the 18 countries surveyed (Figure 1.6). Like in the case of prepaid mobile broadband, Teleco used the price of promotional offers, as long as they did not establish loyalty periods greater than 12 months. Converting Brazilian prices into dollars of 2014 and October 2015, we have 98.86 BRL and 27.00 BRL, respectively. This comparison in BRL suggests that there might have been a mistake in the collection of the ITU data, since it is well off the ratio found for fixed broadband and pre-fixed mobile broadband.

FIGURE 1.6. Postpaid Mobile Broadband Package for mobile phones in USD, October 2015 31.6 25.0

15.1 14.0

12.1 11.8

9.3

8.6

7.0

6.8

6.7

5.6

5.6

5.6

5.6

4.7

4.6

4.2

Source: Teleco (2015b, p. 7).

In short, the prices paid by Brazilians, collected by ITU in 2014 for broad‑ band, measured in USD cents per MB (perhaps the most comparable measure),

34

peter knight

were very high. The worst case is the postpaid mobile broadband (the most expensive of all selected countries). The situation of prepaid mobile broadband is better, but even so, only Argentina has higher prices. For fixed broadband, Brazil ranks better: only China and Argentina had higher prices. Considering the affordability (prices as a percentage of Gross National Income per capita), Brazil’s position was better, but it still left much to be desired. Brazilian prices collected by Teleco in October 2015 for broadband ser‑ vices were lower, measured both in USD and USD PPP. They compared more favorably with those of the other surveyed countries. The reasons for this more favorable outcome seem to be the date of the survey (with a much higher exchange rate than in 2014) and, in the case of mobile broadband, the use of promotional offer prices.

FACTORS AFFECTING THE PRICE OF BROADBAND Why are prices so high in Brazil? There are several factors. The most obvi‑ ous is high taxation (taxes, “contributions” and fees) on telecommunications. Other factors that affect the price of broadband are the lack of competition in some markets, financing and network deployment costs in remote and/or low-income areas, local content requirements in telecommunications equip‑ ment and the high cost of space rental on electric poles, towers and/or fiber optic cables from other operators.

Taxes on telecommunications7 The tax burden in this sector is among the world’s heaviest, 43% in 2015 (Figure 1.7). In 2005, it was already the highest among the eight major sec‑ tors of the economy studied by Rogério Werneck, more than twice the tax burden for the manufacturing sector (WERNECK 2008, tab. 5). A detailed study by Teleco from 2015 examined taxation in different industrial sectors in Brazil and 17 other countries. The study pointed out that Brazil had the high‑ est average tax burden (43%) of the sample countries, far above Argentina, which comes second (26%). The tax burden in Brazil was more than dou‑ ble the average of the other 17 countries (16%). The study found that the tax burden in Brazil was 65% higher than that of Argentina, the second high‑ est among the countries analyzed, 139% than that of Russia and 4.3 times 7

The issue of high taxation on the telecommunications industry is addressed in detail in Chapter 8 of this book.


35

higher than South Korea (Figure 1.7). And in 2016, the Tax on Goods and Services (ICMS), the heaviest tribute, rose an average 3.9% in 11 states and in the Federal District, raising the national average tax burden (weighted by the number of cell phones in each state) from 43% to 46%, and up to 68.5% in the state of Rondônia.8 This tax does not include that of three sector funds mentioned in the next paragraph.

FIGURE 1.7. International comparison of taxation on the telecommunications industry, 2015

Source: Levy (2015).

The exceptionally high taxation of this sector is not only a distortion that re‑ duces the efficiency of investments, it also goes against digital inclusion policies promoted by federal, state and local governments. Moreover, the Ministry of Communications has evidence that a reduction in current rates would increase government revenues because of increased demand for services, more than pro‑ portional to the reduction in prices (KNIGHT, 2014, p. 88-89 and Chapter 8 of this book). From 2001 to 2015, 90 billion BRL in current prices were collected (not adjusted for inflation) by three federal funds, to be used to support the telecom‑ munications industry – the Fund for the Universal Access to Telecommunication Services (Fundo de Universalização dos Serviços de Telecomunicações – Fust), the Telecommunications Inspection Fund (Fundo de Fiscalização das Telecomunicações – Fistel) and the Fund for the Technological Development of Telecommunications

8

See Chapter 8, Table 8.2, in this book.

36

peter knight

(Fundo de Desenvolvimento Tecnólogico das Telecomunicações) Funttel). But out of this total, only 7% were applied to the stated goals of these three funds.9

Other factors affecting the price of broadband Lack of competition among telecommunications companies in some mar‑ kets may be a factor influencing the price of fixed broadband, since many regions and municipalities have only one or a few providers, at least for whole‑ sale. However, the competition is much higher for mobile broadband. There are four major mobile phone carriers in Brazil: Vivo, TIM, Claro and Oi. These companies compete with each other as well as with regional and small‑ er providers, such as Algar and Nextel. In September 2015, 42% of the 5,665 Brazilian municipalities, accounting for 83% of the population, had between four and six carriers competing in the area of Personal Mobile Service (SMP) (TELEBRASIL, 2015, tab. 4.11 and 4.14 p. 105 and 109).10 High financing and network deployment costs are influenced by the high in‑ terest rates prevailing in Brazil (among the world’s highest), high taxes levied on the equipment and cables required as well as the difficulties in reaching remote areas in a huge territory. Domestic content requirements for the equipment and cables used in the construction of networks is another factor. The national content for network equipment and components produced with national technology varies be‑ tween 5 and 30%. For example, it is 30% for the deployment of optical trans‑ port networks, including through optical ground wire cable (OPGW) hanging from the transmission towers of power companies, especially the Eletrobras group, and optical access networks.11 The high cost of renting infrastructure from other carriers (eg. fiber or capacity in fiber optic cables, use of state power company distribution poles to hang fiber optic cables, ducts for cables, space for antennas in towers) is often mentioned as a problem by small and medium-sized carriers. This factor is addressed in the context of the General Plan on Competition Targets (Plano Geral de Metas de Competição – PGMC) by Anatel, discussed in Chapter 4 of this book. 9

For details on these sector funds and their use, see Chapter 8 of this book.

10 For details on the status of competition, see Figures 4.9 and 4.10, Table 7.1 and Chapter 11 of this book. 11 Available at: . Accessed: February 6th. 2016


37

Finally, we can highlight low public investment in networks. In the MultiYear Plan (Plano Plurianual – PPA) from the federal government for the years 2012/2015, investment in the National Broadband Plan (Plano Nacional de Banda Larga – PNBL) was estimated at 2.8 billion BRL in the period 2012/2013. However, budget-control laws for those years allocated only 314.7 million BRL for that. With the curtailment of resources by the Ministry of Finance, the amount was reduced to 267.9 BRL, but the actual investment was only 214.1 million BRL, only 7.6% of the PPA estimate (DINIZ, 2014 p. 17).

Quality In addition to price and transmission rate, it is important to consider oth‑ er elements affecting broadband quality. There are no international statistics on the quality of access. In Brazil, in response to continued complaints about the poor quality of telecommunication services, in 2011 and 2012 Anatel set 14 goals for carriers, including the rate of Data Connection Outage ( 95%) and the Guaranteed Average Transmission rate13 (> 80%).14 For fixed broadband, considering the indicators of all operators moni‑ tored in the first half of 2015, the percentage of compliance with service goals reached only 59.5%. This level was lower than during the years 2012 (70, 94%), 2013 (70.55%) and 2014 (67.85%).15 For mobile telephony, in the first half of 2015, the percentage of compli‑ ance with service goals reached 68.1%. This level was in line with the number for years 2012 (66.97%), 2013 (68.75%) and 2014 (68.78%).16

12 Defined as the median of the speed values from the samples collected at each measurement. Available at: . Accessed: February 6th. 2016 13 Defined as the simple arithmetic mean of the results of Instantaneous Speed measurements, carried out for a month. Available at: . Accessed: February 6th. 2016 14 Available at: . Accessed: December 5th. 2016 15 Available at: . Accessed: December 5th. 2016 16 Available at: . Accessed: December 5th. 2016

38

peter knight

FEDERAL BROADBAND POLICIES IN BRAZIL Since 1989, before the arrival of Internet in Brazil, the federal government has been promoting the growth of broadband, initially through the National Research Network (Rede Nacional de Pesquisa), now National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP). Beginning in 1998, this has been done through privatization in the telecommunications sector and the promotion of competition. Since 2002, a variety of digital in‑ clusion programs have been launched.

Programs of the National Network of Education and Research (RNP)17 Ever since its creation in 1989, the RNP, now a social organization linked to the Ministry of Science, Technology and Innovation (MCTI), has been promoting the expansion of broadband in Brazil. Beginning in 1992, the RNP has offered increasingly faster connections to research and higher education institutions through a network of leased lines provided initially by Embratel and, after the telecommunications privatization in 1998, several other telecom operators. As of 2007, starting in the city of Belém in the northern state of Pará, RNP has been building its own metropolitan fiber optic networks in major cities, in partnership with educational and research institutions, state and local govern‑ ments and various entities of the public and private sectors. The channel for this is the Redecomep Program, with funding from the federal government. In 2011, RNP launched the sixth generation of its national backbone, the Ipê net‑ work. Most of it is the result of a ten-year agreement, brokered by Anatel, with the Oi operator, which granted RNP the use of more than 20,000 km of nation‑ wide fiber optic infrastructure. From 2012 on, RNP started implementing the Veredas Novas (New Paths) program, with the goal of providing connections of 1 Gbps and 100 Mbps for federal education and research institutions (already customers of RNP) in locations outside major urban centers, especially in the interior of the states. Most of these connections are provided by national and re‑ gional telecom operators, but the program also includes collaboration between the RNP, Telebras and state-owned ICT companies. Collaborative programs were launched first in the states of Ceará (Etice)18 and Pará (Prodepa), to meet demand for better connectivity for RNP’s academic customers. 17 The creation and expansion of the RNP are analyzed in Chapter 9. 18 See Chapter 13 for a detailed analysis of Ceara’s Digital Beltway, perhaps the best example of this kind of collaboration.


39

At the end of 2015, the Ipê network (RNP’s backbone) included 27 Points of Presence (PoPs), one in each state, as well as branches serving 1,219 research, health, and higher education institutions around the country, benefiting more than 3.5 million users.19 Its length is about 22,000 km. The metropolitan net‑ works program (Redecomep) currently has 26 networks in operation in the first phase of its operation, and a new network under construction, totaling 1,650 kilometers. The second phase, aimed at going further inland, includes ten networks in operation and three under construction, totaling 330 kilome‑ ters.20 Therefore, the total installed network comprises about 2,000 km of op‑ tical cabling. The newest RNP program, Veredas Novas, already activated about 400 circuits of 100 Mbps and 1 Gbps, with an average of 200 km each, start‑ ing at the point of presence of RNP in the state, totaling about 80,000 km.

Privatization of telecommunications and the promotion of competition The privatization of telecommunications in Brazil dates back to 1998, based on the General Telecommunications Law – (Lei 9742, Lei Geral de Telecomunicações – LGT) of 1997. This law also created the National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel), which regulates telecommunications and promotes competition in the sec‑ tor, matters dealt with in Chapter 3. Chapter 4 examines regulatory policies more broadly, and Chapter 12 details the role of new entrants in the market to promote competition. The end of state monopolies and the promotion of competition among large, medium-sized and small private corporations were fundamental both to expand of broadband infrastructure and to reduce pric‑ es of telecommunication services (although the cost of broadband remains high). Competition was also stimulated by the existence of a large number of small and medium-sized local and regional providers. This wide participation of small and medium-sized providers is a remarkable characteristic of the Brazilian market. There are about 8,000 providers, 4,000 of which are for‑ malized with licenses from Anatel.21

19 Available at: . Accessed: December 5th. 2016 20 Available at: . Accessed: December 5th. 2015. 21 See Chapter 11 for a discussion on the role of small and medium-sized Internet and telecommunications service providers.

40

peter knight

Electronic Government – Citizen Assistance Service (Governo Eletrônico – Serviço de Atendimento ao Cidadão – Gesac) Gesac is the largest digital inclusion program funded by the federal gov‑ ernment. Established in 2002, in the administration of president Fernando Henrique Cardoso, it was widely expanded during the administrations of presidents Luiz Inácio Lula da Silva and Dilma Rousseff. The program offers free broadband Internet connection – land and satellite-based – to remote ed‑ ucation centers, schools, healthcare centers, indigenous villages, border cross‑ ings and quilombos (villages in which escaped slaves settled). Gesac is directed primarily to socially vulnerable communities that have no other means of en‑ tering in the world of information and communication technologies. In 2014, the Ministry of Communications commissioned the connection of 6,898 points of access. The Ministry of Health also joined the program, contract‑ ing another 13,000 points. The project is managed by the Digital Inclusion Secretariat of the Ministry of Communications.22

Broadband in Schools The Broadband in Schools Program (Programa Banda Larga nas Escolas – PBLE) aims to connect all urban public schools to the Internet through tech‑ nologies that provide quality transmission rates and digital services to support public education in Brazil. The PBLE was launched on April 4th, 2008, by the federal government, through Decree n. 6424, which amends the General Plan of Goals for Universalizing Fixed Telephone Service Provided in Public Regime (Plano Geral de Metas para a Universalização do Serviço Telefônico Fixo Comutado Prestado no Regime Público – PGMU) established by Decree n. 4769). With the signing of the Amendment to the Term of Authorization of the Exploitation of Fixed Telephony (Termo Aditivo ao Termo de Autorização de Exploração da Telefonia Fixa), authorized operators traded the requirement to install Telephone Service Stations (Postos de Serviços Telefônicos – PSTs) in the municipalities for the deployment of network infrastructure to support Internet connection at high transmission rates in all Brazilian municipali‑ ties and connection in all urban public schools, with maintenance services free of charge until the year 2025.23 Thus PBLE operates with funds from the 22 Available at: . Accessed: December 7th. 2015. 23 Available at: . Accessed:


41

operators and not from the federal budget. The number schools served in‑ creased from 21,345 in 2008 to 62,925 in 2014, comprising 100% of urban public schools in Brazil (TELEBRASIL, 2015 tab. 2.17, p. 54), though the qual‑ ity of many of these connections is problematic. The PBLE is managed jointly by the Ministry of Education and Anatel, in partnership with the Ministry of Communications, the Ministry of Planning and State and Municipal Departments of Education.

National Broadband Program (Plano Nacional de Banda Larga – PNBL)24 Created by Decree n. 7175 / 2010, the PNBL is an initiative from the federal government that aims at increasing the access to broadband Internet in the coun‑ try, especially in the poorest regions. Coordination of the PNBL was assigned to the Steering Committee of the Digital Inclusion Program (Comitê Gestor do Programa de Inclusão Digital – CGPID), established by Decree n. 6948 of August 25th, 2009. The CGPID was composed of representatives from eight ministries, two secretariats and the Presidential Office. It was chaired by the Minister of the Casa Civil (presidential staff). According to Art. 3 of the Decree. 7175:25 It is up to the CGPID, in addition to the duties provided for in art. 2 of Decree n. 6948, 2009, to manage and monitor the PNBL. It shall also: I – determine the actions, goals and priorities of the PNBL; II – promote and foster partnerships between public and private en‑ tities to achieve the objectives set out in art. 1; III – establish the technical definition of broadband access, for the purposes of PNBL; IV – monitor and evaluate PNBL implementation actions; V – publish an annual report of the actions, goals and outcomes of the PNBL.

The PNBL’s founding document established the Connected Brazil Forum (Fórum Brasil Conectado) to serve as a forum for dialogue and participation in the CGPID, to which it was linked. The following stakeholders were to take part December 7th. 2015. 24 See Chapter 3 of this book for more details on the PNBL. 25 Available at: . Accessed: December 8th. 2015.

42

peter knight

in it: entities representing states and municipalities, the Legislature, representatives of operators, equipment manufacturers, software developers, producers of digital content, representatives of users and civil society entities. CGPID was supposed to help formulate a long-term action plan to deal with strategic ICT issues, help develop legal instruments to implement the strategic plan, and create and disseminate knowledge about ICT. It was also charged with en‑ couraging public debate on policies related to broadband and Internet use and dissemination in Brazil (CGPID, 2010, p. 55-56). In the first term of President Dilma Rousseff, management of the PNBL was transferred to the Telecommunications Secretariat of the Ministry of Communications (MC), through its Department of Broadband. The CGPID was closed, making it difficult to promote cross-ministry coordination of the PNBL. The main objective of the PNBL was to connect 35 million households to the World Wide Web by 2014. The MC worked on several fronts. The most important were the tax relief on networks and access terminals, the Special Taxation Regime of the National Broadband Program (Regime Especial de Tributação do Programa Nacional de Banda Larga – REPNBL),26 the Popular Broadband program and the expansion of the public fiber optic network (managed by Telebras, a federal public sector company reactivated in 2010). The REPNBL-Redes program seeks to encourage investment in net‑ works by the private sector, through exemptions from federal taxes (PIS/ PASEP, Cofins and IPI). From 2013 through 2014, these exemptions added up to 1.6 billion BRL. The Ministry of Finance estimate for 2015 was 1.1 billion BRL. (MINISTÉRIO DA FAZENDA, 2015, p. 28; MINISTÉRIO DA FAZENDA, no date, Chart III). In order to have access to the REPNBL, a company seeking to build networks had to go through several technical and bureaucratic screenings. For example, in 3G and 4G telephony projects, companies had to spend 50% on equipment, 70% of which manufactured according to the a Basic

26 The REPNBL was established by Law n. 12,715 / 2012 and provided exemptions from federal taxes (IPI, PIS/PASEP and Cofins) for broadband infrastructure projects, including fiber optic networks. Available at: . Accessed: December 7th. 2015.


43

Production Process (Processo Produtivo Básico – PPB),27 and 20% of the total expenses on domestic technology. Then the project had to be reviewed by the Ministry of Finance, which granted (or not) the final qualification for the REPNBL (SARDEMBERG, 2013). From 2013 to September 2015, the investments of companies in fixed and mobile services totaled 82.2 billion BRL (TELEBRASIL, 2015, p. 68). The total value of tax relief initiatives in 2013/2015 (2.7 billion BRL) equaled 3.2% of the amount invested in the same period. The Popular Broadband program (Banda Larga Popular) is implement‑ ed via special contracts that offer 1 Mbps Internet connections for 35 BRL per month (with taxes). In 2009, the National Council for Financial Policy (Conselho Nacional de Política Fazendária – Confaz), through its agreement 38/2009, provided for the exemption of the state-level value added tax (Imposto sobre Circulação de Mercadorias e Serviços – ICMS) on basic 1 Mbps broadband contracts. In total, 14 state governments signed the agreement, but only São Paulo, Rio de Janeiro, Espírito Santo, Paraná, Goiás, Pernambuco and Pará regulated the decision through decrees.28 In São Paulo, the first state to exempt popular broadband contracts, in October 2014, 2.5 million homes had this service, paying only 29.80 BRL per month. That month, the governor signed a decree that maintained the reduction, from 25% to zero, for 1 Mbps packages. The decree also extended the tax exemption to two new service groups, covering 1.5 Mbps rate contracts, which may be marketed at 34.90 BRL, and 2 Mbps access packages, at 39.90 BRL.29 By the end of 2014, 32.3 million households had Internet connections, 21.7 million of which were fixed broadband (62% of the PNBL target). Of these, 6.7 million had fixed broadband connections faster than 4 Mbps (CETIC, 2015 tab. A4A6, p. 322-324). In 2014, the Banda Larga Popular package was offered in wholesale in 4,157 municipalities. In retail, it was available in 5,376 munic‑ ipalities, 90% of the 5,565 Brazilian municipalities (TELEBRASIL, 2015, tab. 2.20, p. 58). However, the number of accesses through these packages added 27 A PPB sets the minimum required manufacturing steps that companies must follow to make a given product as one of the counterparts to the tax benefits established by law. 28 Available at: ;. Accessed: October 6, 2016. 29 Available at: ;. Accessed: February 7th 2016.

44

peter knight

up to only 2.6 million in June 2014 (MINISTÉRIO DAS COMUNICAÇÕES, 2014). This suggests that the overwhelming majority of BLP contracts were in the state of São Paulo. Telebras and the expansion of the broadband public network. One of the main objectives of the PNBL was to build a national broadband network, including infrastructure and operations to support the formulation of public policies related to mass access not only to the Internet, but also to official content encouraging social inclusion, civil rights, education and digital culture, among others. (CPID, 2010, p. 41-42)

The responsibility for achieving this goal has been assigned to Telebras. Telebras had the goal of reaching 4,278 municipalities with its fiber optic network by the end of 2014, but the state company said it only reached 612 municipalities, 360 by direct supply and 252 through partners (Diniz, 2014, p. 17). Senator Anibal Diniz, rapporteur of an evaluation of the PNBL for the Federal Senate (Diniz, 2014), wrote in his report that over the first four years of the PNBL Telebras was only able to spend about 284 million BRL, only 7.4% of the planned budget. Moreover, the CGPID had not met since 2010 nor had it presented the PNBL monitoring reports that were its responsibility. The Connected Brazil Forum, created to bring together more than 60 institu‑ tions of governments, civil society and private sector, was deactivated (REIS; FONTENELLE, 2014; DINIZ, 2014, p. 17 and 32). We must also note that the information available on Telebras’ website is very scarce and opaque. The last available analysis by the MC, presented in June 2014, shows Telebras’ coverage data from 2013 and plans for deploy‑ ment in 2014 (Figure 1.8). Most of the Telebras network consists of fiber in OPGW cables, leased from companies in the Eletrobras system. On its website, Telebras states that in early March 2016 its fiber optic network was 28,000 km long.30

30 Available at: ;. Accessed: March 3rd. 2016


45

FIGURE 1.8. Telebras: land network

Ministry of Communications Telebras: Fiber Optic Network December 2013 885 Municipalities covered by the network 269 Municipalities receiving commercial services Deployed in 2012 Deployed in 2013 Deployed in 2014 Agreements and partnerships Unasul Project Segment deployed in 2011

Source: adapted from ministério das comunicações (2014)

Broadband for All (BLPT) and the New Model of Telecommunications31 During the year 2014 there were several official statements and news anal‑ yses of the “PNBL 2.0” proposal – which ended up being called Broadband for All (Banda Larga para Todos – BLPT). It was announced by President Dilma Rousseff during her re-election campaign that year. President Rousseff pro‑ posed universal broadband in the country by 2018 (BOCCHINI, 2014). According to her, universalizing means that 90% of all households should have broadband, via optical fiber with transmission rates of at least 25 Mbps. We believe that the Internet has the same importance as, for exam‑ ple, universal access to electricity. Today, the Internet is as import‑ ant as electricity itself. It is something that we cannot fail to consider as part of people’s lives. It is part of our everyday life. (ROUSSEFF cited BOCCHINI, 2014 [online])

31 See Chapter 3 of this book for further details on the BLPT and the New Telecommunications Model.

46

peter knight

In her speech to the Congress on January 1st, 2015, when she took office for the second time, President Rousseff reiterated this commitment: “I reiter‑ ate my commitment to promoting universal access to the Internet in cheap, fast and safe broadband within the next four years” (BRAZIL, 2015). The year 2015 was marked by debates inside and outside the federal gov‑ ernment (especially in the MC and Anatel, but also in civil society and in the trade press) on the feasibility of achieving these goals and how to finance the program, possibly using some combination of funds from the federal budget, sector funds and private sector investment. By January 2016, the government had not yet determined the goals, funds and program structure. But in November 2015, the MC opened a public consultation on a new telecommunications model, a matter related to the BLPT. By the deadline of January 15th, 2016, it had received 915 contributions from individuals, busi‑ nesses and civil society organizations. These are being examined by a Working Group formed by representatives of the MC and Anatel. The Working Group was supposed to submit a report consolidating the contributions and giving a context of the current market, with international com‑ parisons. Additionally, the group must propose more than one reg‑ ulatory model, explaining the differences between them. The set of proposals was to be evaluated by the Minister and the President so that they could choose a model. Based on this political decision, le‑ gal instruments were to be prepared. These could be bills, decrees, Anatel regulations, among others.32 (ABRANET, 2016 [online])

CONCLUSIONS Throughout this book, the authors deepen the analysis of the aforemen‑ tioned issues. Chapter 16 provides a summary of the main findings and critical factors for the expansion and universal access to quality broadband in Brazil. In this opening chapter we emphasize three topics: (1) the importance of holistic strategic planning for the use of ICT and its analog complements to expedite the country’s economic, social and political development; (2) the evolution of broadband in Brazil compared to other countries; and (3) federal government programs aimed at expanding broadband and digital inclusion. 32 President Rousseff was removed from Office in impeachment proceedings in May of 2016. As of July 2016 her Trial by the Senate had not ended, and the BLPC program was effectively in limbo.


47

Lack of strategic planning There has never been a national ICT strategy to expedite the social, eco‑ nomic and political development of Brazil. The only attempt was the so-called Green Book (TAKAHASHI, 2000) published by the Ministry of Science and Technology. However, it was never implemented. The PNBL, released in 2010, did not include a holistic strategy for the use of ICT. It was limited to the expansion of broadband. The BLPC had not yet been released as of July 2016. There were no signs that it would be integrated into a such a strategy.

Broadband in Brazil and in other countries Despite significant improvements, fixed broadband Internet reached only 34% of households in 2014. Connections of 8 Mbps or higher reached only 12% of all Brazilian homes. All these indicators were worse in rural areas, less developed regions, and among the poor, the less educated and the elderly. The penetration of broadband, the transmission rates, the prices and the quality of broadband connections in Brazil fall short of all expectations. Compared to countries of similar levels of development, with some exceptions, Brazil is not doing a good job.

Federal programs for the expansion of broadband and digital inclusion The expansion of broadband in Brazil was not a priority for former Presidents Fernando Henrique, Lula and Dilma. In the two terms of Fernando Henrique, the priorities were privatization of telecommunications compa‑ nies, establishment of Anatel to regulate these companies and e-government. In Lula’s and Dilma’s terms, e-government was left aside and the emphasis was placed on digital inclusion, largely through obligations imposed on pri‑ vate companies, and then the expansion of broadband as of the launching of the PNBL in 2010, even though it only managed to achieve 62% of its primary goal. Despite the statements of President Dilma, the promised BLPT had not been launched as of July 2016. The federal budget for digital inclusion and expansion of broadband is very low, and these funds are often held back (contingenciados) by the Ministry of Finance (as in the case of Telebras). The Ministry of Finance diverted al‑ most all the Fust revenue and most of Fistel’s and Funttel’s, reallocating them to uses that have nothing to do with the development of telecommunications. These revenues could have been used to promote digital inclusion, make

48

peter knight

public investments and encourage private investment. The tax burden of the sector is one of the highest in the world, with a negative impact on prices and on the ability of the private sector to invest in network expansion. The total of tax exemptions of the REPNBL (called a tax expenditures by economists) were very small in comparison with the huge tax burden imposed on the telecom‑ munications industry.

REFERENCES ABRANET. Consultation on new telecom model received 915 contributions. Available at: . Accessed: February 3rd. 2016. AKAMAI. Akamai’s State of the Internet 3rd Quarter 2015 v. 8, no. 3, 2015. Available at: . Accessed: October 6, 2016. ANATEL. Executive Secretariat. Proposed General Plan on Competition Targets – PGMC. Presentation at Public Hearing, Public Consultation n. 41, July 26th. 2011. Available at: . Accessed: October 6, 2016. BELSON, D. State of the Internet Metrics: What Do They Mean? Available at: . Accessed: December 19th. 2015. BOCCHINI, Bruno. Dilma says that by 2018, broadband in Brazil will be universal. EBC AGÊNCIA BRASIL, 2014. Available at: . Accessed: December 8th. 2015. BORGES. Fundos de Telecom bancam até ferrovia. In: O Estado de S. Paulo, December 10th 2015. Available at: . Accessed: December 19th. 2015. BRAZIL. Speech of President Dilma Rousseff at the time of the Constitutional Commitment, before the National Congress. Department of state, 2015. Available at: . Accessed: December 8th. 2015.


49

BRUNO, L. Programa de banda larga se aproxima do fim criticado por entidades e operadoras. Available at: . Accessed: December 7th. 2015. COMITÊ GESTOR DA INTERNET NO BRASIL (CGI.br). ICT Households and Enterprises – Survey on the use of Information and Communication Technologies in Brazil, volumes for the years 2005-2014. São Paulo: NIC.br e Cetic.br, 20072015. Available at: . Accessed: Dec. 4th, 2015. COMITÊ GESTOR DO PROGRAMA DE INCLUSÃO DIGITAL (CGPID). Secretaria Executiva. Plano Nacional de Banda Larga. Brasília, 2010. Available at: . Accessed: Feb. 7th, 2016. DINIZ, A. Avaliação do Programa Nacional de Banda Larga (PNBL). Comissão de Ciência, Tecnologia, Inovação, Comunicação e Informática (CCT) do Senado Federal, 2 dez. 2014. Available at: . Accessed: Feb 2nd, 2015. DUTTA, S.; GEIGER, T.; LAVIN, B. The Global Information Society Report 2015: ICTs for Inclusive Growth. Geneva: World Economic Fórum, 2015. FERREIRA, M. A arrecadação e a destinação dos fundos das Telecomunicações. Apresentação para Audiência Pública, 17 jun. 2015. Brasília: Ministério das Comunicações. Available at: . Accessed: Dec. 24th, 2015. JENSEN, M. Broadband in Brazil: a Multipronged Public Sector Approach to Digital Inclusion. Washington: InfoDev, 2011. Available at: . Accessed: June 15th, 2015. HANNA, N.; KNIGHT, P. Seeking Transformation Through Information Technology: strategies for Brazil, China, Canada and Sri Lanka. New York: Springer, 2011. ______; ______. National Strategies to Harness Information Technology; Seeking Transformation in Singapore, Finland, the Philippines, and South África. New York: Springer, 2012. HANNA, N.; SUMNER, R. Transforming to a Networked Society: guide for Policy Makers. Gaithersburg, Maryland: Sriban Books, 2015. Available at: . Accessed: Dec 4th, 2015.

50

peter knight

INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA (IBGE). Pesquisa Nacional por Amostra de Domicílios 2014. Rio de Janeiro: IBGE, 2015. Available at: . Accessed: Dec. 7th, 2015. INTERNATIONAL TELECOMMUNICATIONS UNION (ITU). Measuring the Information Society Report 2014. Geneva: ITU, 2014. Available at: . Accessed: June 15th, 2015. ______. Measuring the Information Society Report 2015. Geneva: ITU, 2015. Available at: . Accessed: Dec. 3rd, 2015. ______. Broadband Series. Genebra: ITU, abr. 2012. INTERNET SOCIETY (ISOC). Internet Society Global Internet Report 2014: Open and Sustainable Access for All. Reston, VA e Genebra: ISOC, 2015. Available at: . Accessed: Jun. 16th, 2015. KATZ, R. The Impact of Broadband on the Economy: Research to Date and Policy Issues. Geneva: ITU, 2012. Available at: . Accessed: Apr. 27th, 2016. KNIGHT, P.; FERNANDES, C. (orgs.). e-Brasil: um programa para acelerar o desenvolvimento socioeconômico aproveitando a convergência digital. São Caetano do Sul: Yendis, 2006. KNIGHT, P.; FERNANDES, C.; CUNHA, A. (orgs.). e-Desenvolvimento no Brasil e no mundo: subsídios e programa eBrasil. São Caetano do Sul: Yendis, 2007. KNIGHT, Peter. The Internet in Brazil: Origins, strategy, development and governance. Bloomington, Indiana: AuthorHouse, 2014. LEVY, E. Telecomunicações no Brasil. Apresentação na Audiência Pública na Comissão de Fiscalização Financeira e Controle da Câmara dos Deputados. 17 set. 2015. Available at: . Accessed: Dec. 24th, 2015. MINISTÉRIO DAS COMUNICAÇÕES. Programa Nacional de Banda Larga – PNBL: situação em junho de 2014. Available at: . Accessed: Mach 30th, 2016.


51

MINISTÉRIO DA FAZENDA. Análise da Arrecadação das Receitas Federais, Dezembro/2014. Brasília: Secretaria da Receita Federal, Centro de Estudos Tributários e Aduaneiros, 20 jan. 2015. Available at: . Accessed: Feb 7th, 2016. ______. Demonstrativo dos Gastos Tributários: PLOA 2015. Brasília: Secretaria da Receita Federal, Centro de Estudos Tributários e Aduaneiros, [s.d.]. Available at: . Accessed: Feb 7th, 2016. QIANG, C.; ROSSOTTO, C. Economic Impacts of Broadband. cap. 3. In: WORLD BANK. Information and Communications for Development 2009: Extending Reach and Increasing Impact. Washington, DC: The World Bank, 2009. Available at: . Accessed: May 15th, 2016. REIS, A.; FONTENELLE, A. Banda larga maior depende de recursos e gestão articulada. Jornal do Senado, 2 dez. 2014. Available at: . Accessed: Dec. 7, 2015. SARDEMBERG, C. A. Nos balcões do governo. O Globo, 21 mar. 2013. Available at: . Accessed: Feb 9th, 2016. SINDITELEBRASIL. Setor de Telecomunicações em 2015. Apresentação, Brasília, 24 nov. 2015. Available at: . Accessed: Dec. 24th, 2015. TAKAHASHI, T. (org.). Sociedade da Informação no Brasil: Livro Verde. Brasília: Ministério de Ciência e Tecnologia, 2000. Available at: . Accessed: Dec. 24th, 2015. TELEBRASIL. O Desempenho do Setor de Telecomunicações no Brasil: séries Temporais 9M15. Rio de Janeiro, dezembro de 2015. Available at: . Accessed: Feb 7th, 2016. TELECO. O Desempenho Comparado das Telecomunicações do Brasil: preços dos Serviços de Telecomunicações, Utilização de Banda Larga Fixa. Documento preparado por solicitação da TELEBRASIL & FEBRATEL. São Paulo: Teleco, out.

52

peter knight

2015a. Available at: . Accessed: Feb 6th 2016. ______. O Desempenho Comparado das Telecomunicações do Brasil: preços dos Serviços de Telecomunicações, Utilização de Banda Larga Móvel. Documento preparado por solicitação da TELEBRASIL & FEBRATEL. São Paulo: Teleco, out. 2015b. Available at: . Accessed: Feb 6th, 2016. WERNECK, R. Tax reform in Brazil: an evaluation at the crossroads. Texto para Discussão n. 558, Departamento de Economia, PUCRio, jan. 2008. Available at: . Accessed: Dec. 5th, 2015. WORLD BANK GROUP. World Development Report 2016: Digital Dividends. Washington, DC: The World Bank, 2016. Available at: . Access on: July 23rd, 2016.

|2|

FIXED AND MOBILE BROADBAND: WHAT IS THE FUTURE OF THE MARKET?1 Eduardo Tude [ Teleco ]

T

INTRODUCTION

his chapter presents projections of how broadband will evolve in Brazil during the coming years and discusses the role of government and public policies to expedite this process. Meeting expectations regard‑ ing Internet access necessarily involves the combination of fixed broadband – which has more capacity and supports higher speeds – and mobile broad‑ band, the main advantage of which is enhanced mobility. A country as large as Brazil requires government action to foster the expansion of these offer‑ ings by providers all across the country. The government, however, must re‑ main neutral with regard to technological alternatives. For a country as big and diverse as Brazil, there can be no single solution. The first section introduces some basic concepts, in order to standardize the terminology used in the following sections. The second section shows the current stage of fixed and mobile broadband markets in Brazil. The third investigates the future of these markets. Conclusions and recommendations are presented in the last section.

BASIC CONCEPTS Figure 2.1 summarizes the current profile of Brazilian telecommunications services. We can notice that the trend is the provision of various services (voice, data and video) over the Internet. The connection tends to be made through

1

The figures and charts in this chapter have been developed by the author using statistics available on the author’s own publications and database available on portals like Telebrasil, Teleco and OECD, mentioned in the references at the end of the chapter.

54

eduardo tude

broadband access (fixed or mobile). Many of these services are provided by the so-called Over The Top (OTT) providers, like Netflix or WhatsApp.

FIGURE 2.1. Broadband access in a connected society

Source: Teleco.

In this scenario, smartphones now occupy a central place in people’s lives. They have become the most widespread device for people to connect to the Internet. We now expect to have fast connections anytime, anywhere. Smartphones can connect to the Internet via mobile broadband, mobile net‑ work or fixed broadband via Wi-Fi. This trend is present in Brazil and in the world, though not uniformly for all people. Its dissemination depends on cultural and economic aspects and on the availability of connectivity through broadband access. Connectivity helps to increase productivity in the economy and in people’s personal lives. This process will gain momen‑ tum as billions of objects become connected to the Internet worldwide (The Internet of Things – IoT).

Defining what broadband is There are several definitions for broadband, which makes it difficult to compare statistics from different sources. For the broadband commission of the International Telecommunication Union (ITU) and for UNESCO, broad‑ band is characterized by:

fixed and mobile broadband

55

i. It is always available (always on), with high capacity; ii. It can transport large amounts of data per second and not at a particu‑ lar speed; and iii. It enables the combined delivery of voice, data and video simultaneously. Broadband definitions based on speed tend to become obsolete quickly, due to the growing demand for increasingly higher speeds. The ITU first de‑ fined as broadband subscriptions those with top speed of 256 kbps (down‑ link). The Federal Communications Commission (FCC) of the United States decided in 2015 to call broadband only subscriptions at speeds above 25 Mbps (downlink) and 3 Mbps (uplink). In this chapter, we adopted the definition of broadband considered by the regulator, the National Telecomunications Agency (Agência Nacional de Telecomunicações – Anatel): • Fixed broadband access is made through providers of Multimedia Communication Service (Serviços de Comunicação Multimídia – SCM), regardless of its speed. • Mobile broadband access is through mobile phones with 3G or 4G tech‑ nology (same criteria used by the ITU) and data terminals of any tech‑ nology used for broadband Internet access. It excludes terminals for data communication between objects (machines) of the Internet of Things.

Fixed broadband technologies In fixed broadband, Internet access takes place in a fixed location (office or home) through a pair of copper cables, coaxial cable, optical fiber or radio waves. The first fixed broadband networks were deployed using pairs of cop‑ per wires from fixed telephony providers with Asymmetric Digital Subscriber Line (ADSL) technology. Depending on the quality of the copper pairs, with this technology we can achieve download speeds between 1 and 2 Mbps for a distance of up to 4 km from the network wire concentration point, usually where the central office is located. The development of new technologies2 in the xDSL family (ADSL, ADSL2, HDSL, VDSL) and – more recently – Vectoring and G.Fast enabled the deployment of broadband access using copper cables with high speeds (above 100 Mbps), but for short distances (less than 500 meters). This led 2

xDSL is a family of broadband access technologies that uses the fixed telephony subscriber’s copper wire as the transmission medium (Digital Subscriber Loop).

56

eduardo tude

many providers to adopt the architecture in which the fiber goes to a cabinet (network node) in the customer’s neighborhood and the last meters of the access are made through copper. This broadband solution is called fiber to the node (FTTN). New networks are increasingly adopting the architecture in which fiber goes all the way to the home (Fiber to the Home – FTTH), building entrance (Fiber to the Building – FTTB) or to the node in the neighborhood (Fiber to the node – FTTN). Fiber permits very high speeds regardless of the distance involved. Furthermore, installation costs for fiber connections are increasingly similar to those for other solutions using copper wires or coaxial cable. At any rate, the use of existing copper networks can still be a good option for the internal network of the a house or building, given the short distances involved and the high cost of putting new cables in these locations. The set of fiber solutions in broadband access is called FTTx (FTTN, FTTB, FTTH) and is the basis for the expansion of high speed fixed broad‑ band. In Brazil, it has been adopted not only by large operators (Vivo, GVT, Oi and TIM), but also by small and medium-sized providers. There is also the hybrid architecture of fiber and cable (HFC) used by cable operators such as NET in Brazil. With the increased speed of connections, the trend is that fiber participation in hybrid solutions (FTTN, HFC etc.) will grow and get closer and closer to the subscriber. Optical fiber has high capacity and no limits to increases in bandwidth speed. However, fiber-based solutions might not be economically viable in scarcely populated areas. For these regions, there is the option of fixed wireless broadband via radio, satellite or cell phone network (3G/4G). A solution used by small and medium-sized providers is radio with spread spectrum technology, in frequency bands that do not require any license from Anatel. In fixed broadband 4G networks, a network similar to mobile broadband is used, but the user mobility is restricted to a small area near the installation site.

Mobile broadband technologies In mobile broadband, users have mobility because access is through mo‑ bile networks. It is done through radio waves between the smartphone and the Radio Base Station (RBS, ERB in Portuguese), popularly known as cell


57

tower. Table 2.1 shows download speeds for smartphones according to the technology of cellular networks.

TABLE 2.1. Download speed per technology in Mbps Mbps

3G

3G+

4G

Advanced 4G

Maximum

14.4

42

100

1,000

Typical per user

0.3-1

3-6

5-12

40

Source: Teleco.

Top speeds for each technology are obtained in the RBS and shared by multiple users using smartphones, tablets or other devices. Usually, the pro‑ vider limits the maximum speed offered to the individual user at a rate close to the typical speed. Base stations are connected to the core of the cell phone network. In order to support the traffic generated with 3G+/4G technolo‑ gies, such high-speed connections have to be made mainly through optical fiber. That is, the optical fiber network that supports fixed high-speed broad‑ band can also be used to connect cell towers. Hence it can also support the provision of mobile broadband.

Backhaul Backhaul is the portion of a hierarchical telecommunications network re‑ sponsible for making the connection between the core network or backbone and remote subnets. Thus, the network that connects the base stations to the core of the mobile network is called backhaul for the mobile network. The net‑ work connecting municipalities to the national network of a provider (back‑ bone) is also called backhaul. Note that for the provision of fixed and mobile high-speed broadband in a given municipality, having fiber and/or 3G/4G-based connection is not enough. The municipality must have a high-speed connection, preferably by optical fiber, with the national backbone of a provider.

Mobile vs. fixed broadband Table 2.2 presents the advantages of mobile broadband and fixed broad‑ band networks:

58

eduardo tude

TABLE 2.2. Advantages of each type of broadband Mobile broadband

Fixed broadband

Mobility

Speed

Deployment cost

Capacity

Entry price

Price per byte

Source: Author’s survey.

The big advantage of mobile broadband is mobility. It is available any‑ where a 3G/4G cellular network is available. In mobile broadband, the capac‑ ity is shared and data plans have a quota of consumption. Fiber-based fixed broadband networks offer greater capacity and speed for users. Operators can thus provide unlimited usage plans and a lower price per byte. Purely cop‑ per-based networks at distances exceeding 1 km offer speeds similar to those obtained with 3G, but they leg behind fiber networks or 4G networks. Fiber-based fixed broadband networks have a higher implementation cost and therefore are economically viable only in densely populated regions. On the other hand, mobile broadband networks have a lower deployment cost. Thus, fixed wireless broadband with 3G/4G technology can be a good alterna‑ tive to cover areas with lower population density. Overall, residential fixed broadband requires the payment of a fixed month‑ ly fee without any mobility. In mobile broadband, however, there is more flex‑ ibility in payment because of the existence of pre-paid plans. Because of the dynamics described here, for users who have enough income to pay for the service, the ideal is to have high-speed fixed broadband at home and mobile broadband for their cell phones. For low-income users, who have to choose between these two alternatives, mobile broadband will probably prevail.

CURRENT LANDSCAPE OF THE MOBILE AND FIXED BROADBAND MARKET IN BRAZIL Figure 2.2 shows the evolution of fixed and mobile broadband since 2007 and 2008 respectively. Brazil ended 2015 with 217 million broadband access‑ es, 25.6 million of which were fixed broadband subscriptions (at home, office) and 180.5 million were mobile broadband subscriptions (smartphones via cel‑ lular network).


59

FIGURE 2.2. Evolution of fixed and mobile broadband in Brazil (million accesses) 200

180.5

180

157.9

160

Millions

140 120

103.1

100 80

59.2

60

41.1

40 20

10.9

8.3 2.8

0 2007

2008

12.8

20.6

17

19.8

22.2

24

25.6

2011

2012

2013

2014

2015

8.7 15.3 2009

2010

Fixed broadband

Mobile broadband Source: Telebrasil (database).

Fixed broadband At the end of 2015, there were 25.6 million fixed broadband subscriptions in Brazil. 66.4% of this total had speeds higher than 2 Mbps. Dividing these total subscriptions (25.6 million) by the population, we find a density of 12.5 subscriptions/100 inhabitants (TELEBRASIL). Figure 2.3 shows the distribu‑ tion of the density and the percentage of connections with speeds higher than 2 Mbps by states (and the Federal District) in 2015.

60

eduardo tude

FIGURE 2.3. Fixed broadband per state, 2015

up to 6 6 to 12 12 to 18 above 18

up to 50% 50% to 55% 55% to 66% above 66%

Source: Telebrasil.

States in the North and Northeast regions have a lower density of fixed broadband subscriptions (less than 6.0 subscriptions/100 inhabitants). Southern and Southeastern states have densities higher than the average in Brazil. This distribution is not the same when we consider connections faster than 2 Mbps in total fixed broadband subscriptions in each state. The states with low subscription density in the North and in the Northeast have a high percentage of connections with speeds higher than 2 Mbps. This is because most fixed broadband subscriptions in these states are concentrated in their capitals. In Amazonas, for example, 97.8% of fixed broadband subscriptions are in Manaus, the state capital. Fixed broadband subscriptions with xDSL technology (including FTTN) represented 51.8% of Brazilian subscriptions in 2015. Cable TV (Cable mo‑ dem/HFC): 32.3%. Wireless (mainly radio with spread spectrum): 6.9%. Fiber (FTTH/FTTB): 5%. Other technologies: 4%. The distribution is very similar to the average of the Organization for Economic Cooperation and Development (OECD), except with regard to fiber-based subscriptions (Figure 2.4).


61

FIGURE 2.4. Distribution of fixed broadband subscriptions by technology 51.8%

47.6% 32.3% 32.6% 17.1% 6.9%

5% DSL

Cable TV

Fiber Brazil

4%

1.4% Fixed wireless

1.3%

Other technologies

OECD countries Source: OECD Portal and Telebrasil Portal.

The most critical situation for the spread of broadband is found in the 4,439 Brazilian municipalities with less than 30,000 inhabitants (Table 2.3). In 2015, these municipalities held 23.1% of the population, but only 7.7% of fixed broadband subscriptions in Brazil. There were 4.2 subscriptions per 100 inhabitants, and only 34.4% of those connections were faster than 2 Mbps.

TABLE 2.3. Fixed broadband by city population range (2015) Municipalities (inhab.) Up to 30,000 30,000 to 100,000

Acesses fixed BB

Percentage of Subscriptions/ population 100 inhab.

Speed> 2Mbps

7.71%

23.08%

4.2

34.40%

12.07%

20.86%

7.2

42.79%

100,000 to 500,000

30.55%

26.13%

14.6

66.69%

Over 500,000

49.67%

29.92%

20.8

76.85%

100%

100%

12.5

66.36%

Brazil Source: Anatel and Teleco.

In smaller towns there are fewer providers and a higher concentration of subscriptions provided by the fixed telephony provider (67% of fixed broad‑ band subscriptions in those municipalities), mainly using ADSL technology. The rest is provided by small providers. Most of them used radio to provide

62

eduardo tude

their services, but they are already switching to fiber solutions. Wireless tech‑ nology accounts for 25% of all subscriptions in these municipalities.

TABLE 2.4. Market share of fixed telephony operators (2015) Municipalities (inhab.)

Fixed broadband

Up to 30,000

66.8%

30,000 to 100,000

68.8%

100,000 to 500,000

43.4%

Over 500,000

35.6%

Source: Anatel and Teleco.

As shown next, mobile broadband is also scarce in smaller municipalities.

Mobile broadband Brazil ended 2015 with 180 million mobile broadband accesses: 149 mil‑ lion via 3G devices, 25 million via 4G devices, and 6.0 million broadband data terminals (modems). In 2014, 3G surpassed 2G and became the country’s leading mobile technology (Figure 2.5).

FIGURE 2.5. Mobile subscriptions by technology

Millions

Cell phone technology in Brazil

134

157

178

200

196 160

95

2007

145

149

113

66

95

2

4

2008

2009

15 2010 2G (GSM)

52

33

2011

2012 3G

1

7

2013

2014

25 2015

4G Source: Anatel and Teleco.


63

The migration to 3G/4G comes with the growth of smartphone sales. In 2015, these sales accounted for more than 90% of all mobile phones sold in Brazil.3 As shown in Table 2.5, in 2015 3G coverage was available to 94.9% of the population, whereas 4G coverage was available to 55.0% of the population.

TABLE 2.5. Cell phone coverage by technology (2015) Operators/ carriers

Municipalities

% Population covered

2G

3G

4G

2G

3G

4G

Vivo

3,757

3,524

411

91.2%

88.8%

51.5%

Claro

3,814

2,332

187

92.8%

82.5%

48.0%

TIM

3,526

1,741

183

91.6%

75.2%

46.6%

Oi

3,399

1,197

147

88.8%

72.2%

45.7%

Nextel

410

410

10

47.2%

47.2%

5.1%

Algar

106

35

-

2.1%

1.3%

-

2

2

-

0.3%

0.3%

-

5,565

4,420

469

100.0%

94.9%

55%

Sercomtel Total Source: Teleco.

The biggest gap in municipalities served by mobile broadband 3G/4G is among those with less than 30,000 inhabitants, as shown in Table 2.6.

TABLE 2.6. Cell phone coverage by city population (2015) Municipalities (inhab.)

2G

3G

4G

Up to 30,000

100.0%

74.9%

4.1%

30,000 to 100,000

100.0%

95.3%

13.4%

100,000 to 500,000

100.0%

100.0%

68.8%

Over 500,000

100.0%

100.0%

100.0%

Brazil

100.0%

79.4%

8.4%

Source: Teleco.

Broadband offer vs. backhaul offer As mentioned above, to offer broadband in a given municipality requires a high-speed data connection (backhaul) to the national network (backbone) 3

Available at: . Accessed: May 10th, 2016.

64

eduardo tude

of one of the providers. A major breakthrough in expanding backhaul for all municipalities was achieved in 2008, with the issuance of Decree 6424 of April 4th, 2008. It amended the General Plan on Universal Service (Plano Geral de Metas de Universalização – PGMU) for landline operators. The decree estab‑ lished the obligation for such utilities to deploy backhaul in 3,445 municipalities where this connection did not yet exist. In return, the utilities were exempted from the requirement to deploy Telecommunication Service Stations (Postos de Serviços de Telecomunicaçãoes – PST) originally provided for in the PGMU. These new high-speed data connections (backhaul) were implemented via ra‑ dio in 2,363 municipalities, fiber in 902, and satellite in 159. Other technologies were used in 20 municipalities. Figure 2.6 shows the technology distribution in the municipalities when these goals were met by the utilities in 2012.

FIGURE 2.6. Technologies used in backhaul, 2012

Source: Teleco and Anatel (database).


65

We can notice that among the 2,363 municipalities that deployed back‑ haul via radio, 92% have a population smaller than 30,000 inhabitants. Once again, this shows the gap for the provision of high-speed broadband in these municipalities.

THE FUTURE OF THE MARKET Figure 2.7 shows Teleco’s projections of market developments over the 10 years following 2014 if the current regulatory and macroeconomic trends are maintained. In 2014, 40% of mobile phones in Brazil used 2G technology. The migration of these cell phones to 3G/4G will lead to faster growth of mo‑ bile broadband in the next three years. The trend is that by 2020, more than 90% of mobile phones in Brazil will be 3G/4G (mobile broadband). These projections did not consider mobile access to the Internet of Things.

FIGURE 2.7. Projections made by Teleco (million subscriptions) Mobile broadband

Fixed broadband

300 250

Millions

200

158

180

190

200

225

230

235

210

220

240

215

32

34

36

38

40

42

44

150 100 50

24

26

28

30

0 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Source: Teleco.

We expect that the trends observed for mobile and fixed voice services will be repeated in the case of broadband. In 2015, there were about six mobile phones for every fixed phone in Brazil. In 2024, there should be six subscrip‑ tions of mobile broadband for every fixed broadband access. The trend is the

66

eduardo tude

migration of 2G cell phones to cell phones with mobile broadband (3G, 4G and 5G in the future). In 2024, 2G cell phones should account for less than 10% of all mobile phones in Brazil. But we should highlight that, unlike voice services, that enable users to do without landline phones, meeting users’ Internet-related expectations nec‑ essarily involves the combination of fixed broadband, with more capacity and higher speeds, and mobile broadband, with its mobility. Broadband subscriptions should continue to grow slowly, up to 44 mil‑ lion by 2024. If current trends are maintained, it would take 10 years for the density of fixed broadband subscriptions to grow from 12.1/100 inhabitants in 2014 to the current level of fixed telephony accesses, that is 22.7/100 in‑ habitants. The speed of access should continue to grow wherever fiber-based solutions are deployed. The provision of high-speed and affordable access in municipalities with less than 30,000 inhabitants will remain limited until these municipalities have fiber-based backhaul. Brazil has more than 5,000 fixed broadband providers with SCM licenses granted by Anatel. Most of these providers serve the corporate market or a small number of municipalities. They do not need to meet coverage require‑ ments and the expansion of their networks into new geographical areas is subject to the return over their investment. Mobile broadband providers, on the other hand, use a scarce resource (electromagnetic spectrum), with frequency bands periodically auctioned by Anatel. These tenders allow Anatel to establish coverage requirements for the providers that acquire these frequency bands. The 2.5 GHz call for tenders held by Anatel in 2012 established coverage requirements with: • 4G in all municipalities with more than 30,000 inhabitants by 2017, i.e., at least 1,136 municipalities and 76.8% of the population. • 3G in all Brazilian municipalities, including those with less than 30,000 inhabitants, by 2019. However, in the 700 MHz frequencies tender held by Anatel in 2014, the federal government decided to prioritize revenue generation for the Treasury, and the call for tenders did not set service goals. We lost the opportunity to make a commitment to expanding 4G in municipalities with less than 30,000 inhabitants, which would also stimulate the construction of fiber backhaul for these municipalities and create the possibility of fixed 4G broadband.


67

In addition to these smaller municipalities, Brazil has other coverage gaps in other municipalities. The population has pushed the government to ex‑ pand mobile coverage to places like: • Neighborhoods, districts and places distant from the urban centers; • Highways; and • Rural areas. Commitments for expansion of cellular coverage for these sites could have also been included in the tender for frequencies in the 700 MHz band. Finally, it is important to promote the use of new technological advances for providing fixed broadband via satellite (Ka band), which should be available in Brazil in 2016. It should be an alternative to the expansion of broadband to remote areas. The country must also keep up with the constant evolution of mobile technologies (Advanced LTE and 5G), which should offer increasingly high speeds over the next few years. These gaps in the markets should be addressed by public policies aimed at making high-speed broadband accessible to all Brazilians.

CONCLUSIONS AND RECOMMENDATIONS The expansion of fixed broadband (with its capacity and speed) and mo‑ bile broadband (with its mobility) is essential for us to guarantee “access to the world wide web (Internet) for all citizens”, as per the goals set in Decree 4,733 of June 10th, 2003. In a country as big as Brazil, the government should develop an action plan to encourage expansion of broadband for all municipalities. This plan should be carried out by the providers and initially prioritize the construc‑ tion of optical networks (backhaul) for all Brazilian municipalities, encourag‑ ing dark fiber sharing between service providers. The connection to the fiber backhaul enables the provision of: • Mobile broadband (3G, 4G…). • Fixed wireless broadband (4G, for example). • Fixed broadband by small providers with the use of various technologies. This approach would promote competition in these municipalities, re‑ ducing prices and improving service quality. As shown in this chapter, special emphasis should be given to municipalities with less than 30,000

68

eduardo tude

inhabitants, where most connections occur via radio. Reverse auctions can be a good tool to enable these connections. Reverse auctions have been used in several countries, such as Chile and the United States, and they minimize the subsidy given for goals to be achieved. A successful case in Brazil was the Minas Comunica program, which promoted reverse auctions to take cellular service to all municipalities in the state of Minas Gerais. The General Telecommunications Law (Lei Geral de Telecomunicações – LGT) establishes the federal government’s responsibility to organize the op‑ eration of telecommunications services. It is the government’s duty to guar‑ antee access to telecommunications for everyone, to encourage increased use of telecommunications networks and services and to create investment op‑ portunities. The government should act to encourage the provision of broad‑ band Internet services in Brazil by promoting competition and investment. However, the government has done more to drain the financial resources from this sector than to promote investment. This becomes clear analyzing high tax burden in this sector. In this landscape of little public investment, one must assume that the development of broadband in Brazil will take place at private expense. But it is up to the government to create the conditions for this to occur. The govern‑ ment has several instruments to develop public policies, namely: • Reducing the tax burden. • Reviewing the concession model for fixed telephony. • Tendering for other frequencies.

Tax burden4 Of the total paid by users in their telecommunications services bill, at least 30.15% is collected by the government: • At least 25% for the value added tax imposed by the states (ICMS), the same rate applicable to alcoholic beverages, tobacco, perfumes and golf clubs. In some states, the rate reaches 37%. • 3.65% of PIS/Cofins for the federal government. • 1% for the Fund for Universalization of Telecommunication Services (Fundo de Universalização dos Serviços de Telecomunicações – Fust). By 4

The tax burden on the telecommunications industry is addressed in detail in Chapter 8 of this book.


69

June, 2015, 19.4 billion BRL had been collected, but nothing had been invested in the sector. • 0.5% for the Fund for the Technological Development of Telecommunications (Fundo para o Desenvolvimento Tecnológico das Telecomunicações – Funttel). By June, 2015, 5.3 billion BRL had already been raised, but only 30% of this total had been invested in the sector. A study by the Global System Mobile Association (GSMA) found that Brazil has the third highest tax burden for telecommunications services among 50 developing countries, only lower than Turkey and Uganda.5 A study by Teleco conducted in 2015 showed that Brazil had the highest tax burden among 18 countries selected for their size and relevance to the telecommuni‑ cations industry in the world. Together, these countries account for 55.4% of the world’s population (Teleco, 2015, pic. 5.1, p. 7). This means the collection of about 60 billion BRL every year. The gross rev‑ enue of the telecommunications sector accounted for 4.0% of Brazil’s GDP in 2015, while the ICMS tax on communications accounted for 8.6% of all ICMS revenue in the country (TELEBRASIL). Moreover, the government collects about 3 billion BRL every year with inspection fees (Fistel). Anatel’s annual budget ac‑ counts for 11% of this value.6 We can add 52 billion BRL to this figure. Such value has been raised since 1998, with the frequencies tender for mobile phone ser‑ vices. In the 700 MHz tender alone, held in 2014, 5 billion BRL were collected.7 Such a high tax burden has a direct impact on the profits and the invest‑ ment capacity of providers. They invested 30.9 billion BRL in 2014 (Capex), which corresponded to 23.1% of their total net revenue, more than the interna‑ tional average of 15%. The average margin of Ebitda8 for telecom providers in 5

GSMA represents the interests of almost 800 mobile operators worldwide, with over 250 companies in the largest mobile system, including companies that produce mobile phones, software, equipment and Internet services, as well as organizations of associated industries. Available at: . Accessed: February 20th, 2016.

6


7


8

Earnings Before Interest, Taxes, Depreciation and Amortization (EBITDA). That is, the amount of earnings a company generates from its activities without taking into account financial and tax effects. Infomoney. Available at: . Accessed: April 15th, 2016.

70

eduardo tude

Brazil in 2014 was 26.3%, lower than the international average, which is over 30% (Teleco, 2015, tables 2.30 and 2.31, p. 68 and 69; Teleco Portal). There are some positive tax relief initiatives, such as the Popular Broadband (Banda Larga Popular) program, the Special Tax Regime for Telecommunications Networks (Regime Especial de Tributação do Programa Nacional da Banda Larga – REPNBL), and that applied to smartphones. However, these initiatives generate little tax reduction when compared to the figures shown at the beginning of this section. The following initiatives could lead to a gradual reduction of the tax bur‑ den on broadband services, helping to increase investment in the sector and to make services more accessible to low-income consumers: • Setting 25% (still a high value) as the maximum rate of ICMS to be levied by the states for fixed and mobile broadband services. • Exemption from all federal and state taxes for mass (popular) broad‑ band plans for the low-income population served by government so‑ cial programs, as in the United States and other countries. • Exemption from Fust for providers that submit investment projects in broadband networks. • Reduction of inspection fees collected for 3G or higher mobile devices 5.68 BRL (TFF) and 1.89 BRL (TFI). The same values should apply to the Internet of Things. • Exemption from IPI, PIS, Cofins and ICMS taxes9 for all broadband net‑ work equipment manufactured in Brazil for a period of, say, five years. Applicable to companies that invest a minimum percentage, for example, 3% of their net revenues, in technological research and development.

Reviewing the concession model for fixed telephony. Fixed telephony was the main telecommunications service in Brazil and in the world in 1997, when the LGT was passed. This law established a new regulatory framework for the country. According to Anatel, at that time there were 17 million landline phones, 4.6 million mobile phones and 2.5 million pay TV subscriptions in the country. Then there was no residential broadband (Internet service was dial-up) and mobile telephony seen as an expensive ser‑ vice that would not replace the landline. 9

See Table 8.2 in Chapter 8 for details on these taxes.


71

The demand for fixed telephony was high and the waiting list for a fixed phone was only eliminated with privatization, under rules established in the LGT. The landline was seen as basic infrastructure and it should be present in all homes, just like electricity and water. It made perfect sense at the time to establish the PGMU to be carried out by fixed line operators, with goals such as: • Landlines (individual access) in localities with more than 300 inhabi‑ tants. Requests should be addressed within 7 days. • Public Use Terminal (Terminal de Uso Público – TUP), or payphone, in locations with more than 100 inhabitants. The fulfillment of these goals required heavy investments from the con‑ cessionaires (75 billion BRL between 1998 and 2002, with a peak of 24.5 BRL in 2001). From 2001 on, they were able to serve more than 44 million phones. But the rate of utilization of this capacity, which was 90.1% in 2003, dropped to 64.2% ten years later (Figure 2.8).

FIGURE 2.8. Landlines from concessionaires 42.5

42.2

42.3

42.4

42.1

36.8

37.9

38.1

38.0

37.6

42

42.1

43.4

43.7

44.1

44.1

44.9

44.4

35.7

35.1

34.6

33.4

32.1

30.7

29.9

28.5

Millions

35.1 27.8 18.8 17.0

22.1

30.6 25.0

27.2

25.5

20.0

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Installed

In service

Source: Telebrasil (database).

The growth of companies authorized by Anatel to provide local fixed phone service – without having to meet the goals set by PGMU – partly ex‑ plains the decline in landlines from the incumbent operators. In 2015, these concessionaires had 25.5 million fixed telephones, whereas authorized com‑ panies (mainly Embratel/Net and GVT) had 17.8 million. The total number of fixed telephones in service grew by only 9.5% over a period of ten years (2005/2015). Attempts to popularize phones with more affordable prices

72

eduardo tude

low-income people, such as the Special Class Individual Subscription (Acesso Individual Classe Especial – AICE), were not successful. With payphones the trend was the same. In 2005, there were 1.3 million payphones and, ten years later, only 900,000. The average monthly net rev‑ enue per payphone (TUP) fell from 44.0 BRL in 2010 to 10% of this amount in 2013, without taking inflation into account. What we could not foresee in 1997 was that cell phones would become the primary device for voice communications, replacing the landline. The penetration of landlines in Brazilian households fell from 52.9% in 2002 to 37.2% in 2014, while cell phone penetration increased from 34.7% to 88.8% in this period. Promoting access to high speed Internet for all Brazilians is undoubtedly a very important goal for the country. But associating goals with specific tech‑ nologies (mobile or fixed) can lead to distortions, like in fixed telephony. In this process, the rule should be government neutrality in regard to technolog‑ ical alternatives. For a country as big and diverse as Brazil, there are no silver bullets. The goal should be the expansion of the service, without determining the technologies to be used. There is also a clear need for a review of the fixed line concession mod‑ el. The best option would be to plan the gradual end of these concessions, enabling increased investment in broadband. The obligations for fixed line operators only make sense in small towns and remote locations, where the concessionaire operates as a monopoly. A realistic analysis of the issue of revertible assets10 should be part of the incentive for providers, allowing them to invest more in the construction of broadband networks, which is vital infra‑ structure for the growth of Brazil. The establishment of a plan for the gradual end of the fixed telephony concessions in 2025 would enable the transition to a scenario in which broad‑ band will become the leading telecommunications service.

Tendering frequencies The growth of mobile broadband in Brazil depends on Anatel conducting auctions for new frequency bands, which can increase the capacity and speed of today’s networks. It is also important to tender frequencies for fixed wire‑ less broadband with 4G technology to be provided by small Internet providers. 10 Revertible in Brazilian terminology, means that an asset reverts to state ownership at the end of the concession period.


73

In the next frequency auctions, Anatel should no longer aim at raising revenue for the Treasury. It should include expansion commitments in the area served by the mobile broadband – like the 4G service – for: • Municipalities with less than 30,000 inhabitants; • Neighborhoods, districts and places distant from the urban centers; • Roads. In short, to achieve Brazil-wide broadband (fixed and mobile) service, the government must encourage investments and reduce the drainage of funds from this sector.

REFERENCES AGÊNCIA NACIONAL DE TELECOMUNICAÇÕES (Anatel). Available at: . Accessed: February 2nd, 2016. FIBER TO THE HOME COUNCIL EUROPE. FTTHHandbook_2014V6.0. Available in: . Accessed: February 2nd, 2016. OECD BROADBAND PORTAL. Available at: . Accessed: February 2nd, 2016. TELECO PORTAL. Available at: . Accessed: April 3rd, 2016. TELECO. O Desempenho do Setor de Telecomunicações – Time Series Q3 2015, dec. 2015. Available at: . Accessed: February 2nd, 2016. ______. O Desempenho Comparado das Telecomunicações do Brasil, Preços dos Serviços de Telecomunicações: utilização de Banda Larga Fixa Sao Paulo: Teleco, may. 2015. Available at: . Accessed: February 21st, 2016. TELEBRASIL. Database prepared by Teleco. Available at: . Accessed: February 2nd, 2016. TUDE, E. et al. Relatório Telecom no Brasil 2015. São Paulo: Teleco, 2015. ______. Relatório Celular no Brasil 2015. São Paulo: Teleco, 2015.

POLicies and REGULAtion

|3|

NATIONAL BROADBAND PROGRAM AND BROADBAND FOR ALL: CIVIL SOCIETY’S PERSPECTIVE Veridiana Alimonti [ Coletivo Brasil de Comunicação Social – Intervozes ]

I

INTRODUCTION

n the first decade of the 2000s, various countries launched plans to ex‑ pand access to broadband Internet (Intervozes, 2012, p. 86-108). They did so because of the broadband potential to strengthen rights like educa‑ tion, culture, freedom of speech and access to information, and also because of economic reasons, directing investment to creating jobs and stimulating the domestic economy.1 In this context and based on a critical analysis of the Brazilian situation in terms of service availability, price of access and connec‑ tion speed (CGPID, 2010, p. 11-16), the Brazilian government launched the National Broadband Program (Programa Nacional de Banda Larga – PNBL) in May 2010. This initiative reflected recognition that the market alone would not be able to properly expand access to Information and Communication Technologies (ICT) without deepening inequality, especially when it comes to access to high-speed Internet. Thus the program sought to coordinate state action in different dimensions, involving regulation and infrastructure standards, tax incentives, production and technology policies, activation of a national network by Telebras and encouraging the production of content and digital applications. The program also provided for an institutional space for social participation – the Connected Brazil Forum ( Fórum Brasil Conectado) – designed to bring together representatives of companies, uni‑ versities, government and civil society.

1

Some of these countries were the United States, Canada, UK, France, Finland, Australia, Japan, South Korea and Argentina (SOUSA, 2011, p. 9).

78

veridiana Alimonti

In 2011, during the review of fixed telephony concession contracts – held every five years –, the Ministry of Communications and the National Telecommunications Agency (Anatel) negotiated the offering of “mass broad‑ band” (Banda Larga Popular) plans with the providers. Fixed line operators (Telefonica, Oi, CTBC and Sercomtel) committed themselves to providing, by the end of 2014, a 1 Mbps plan for 35.00 BRL (or 29.90 BRL when there is ICMS exemption) in all municipalities where there was land infrastructure for a telecommunications network.2 In addition to this retail offer, the agreement also provided for the whole‑ sale offering of access to the operator’s IP network by other broadband service providers, establishing capacity and link prices. This measure was intended to increase competition in order to ensure easier access of small and medium providers to large carrier networks. The reactivation of Telebras played a sim‑ ilar role in the PNBL: it was assigned the task of building a national network, based on the existing fiber optic infrastructure, linked to federal companies in the electric and oil industries. Together with the wholesale capacity, art. 4th of the PNBL Decree (Federal Decree 7,175/2010) also designated Telebras to implement the private com‑ munication network of the federal public administration and to support pub‑ lic policies for broadband Internet connections in universities, research cen‑ ters, schools, hospitals, service stations, community centers and other places of public interest. Rural areas were not included in the commitment to supply basic Banda Larga Popular subscriptions signed in 2011. These areas were later included in the auction of the 450 MHz band, which occurred in 2012, together with the bidding of the 2.5 GHz band for the provision of 4G mobile service. The objective of this auction was to provide telecommunications services to areas within 30 km from the center of the municipalities, according to the schedule provided in the announcement. Also in 2012, Federal Law 12,715/2012 was approved, giving legal support to the Special Regime of Taxation of the National Broadband Program (Regime Especial de Tributação do Programa Nacional de Banda Larga – REPNBL). 2

The ICMS (Imposto sobre Circulação de Mercadorias e Serviços) is a value-added tax imposed by the states. It makes up the largest part of the high tax burden imposed on telecommunications services (see Chapter 8). This agreement was signed through Commitment Terms that can be accessed at . Accessed: July 2nd, 2015.

national broadband program and broadband for all

79

Regulated in 2013, the initiative provides exemptions of federal taxes for the construction of broadband Internet networks, including taxes like IPI, PIS/ PASEP and Cofins. Five years after its launching, the overall assessment of the PNBL by orga‑ nizations and activists accompanying digital inclusion policies in the country is quite negative. The data and reasons that support this evaluation are pre‑ sented in this chapter. We will also address relevant aspects of new initia‑ tives already announced by the federal government in this area, such as the Broadband for All (Banda Larga para Todos) program, and the current regula‑ tory framework review in the telecommunications industry.

A CRITICAL PORTRAIT OF BROADBAND ACCESS IN BRAZIL If on the one hand accesses to fixed broadband in Brazil increased from 12.9 million in May 2010 (MINISTRY OF COMMUNICATIONS, 2015; TELEBRASIL, 2015a) to about 25.6 million at the end of December 2015,3 on the other hand the initial goal of the PNBL to connect 35 million households by December 2014 was not met. In addition to the gap of 10 million households, the total number at the end of 2014 includes not only households, but also businesses, which correspond to a significant portion of these connections. Therefore, the number of connections in households fell far short of expectations.4 By the end of 2014, the access density of fixed broadband in the country was 11.5 accesses per 100 inhabitants,5 with considerable internal disparities. Such disparities can be seen in Table 3.1.

3

Data consolidated by Teleco Consultoria, based on statistics provided by Anatel. Available on: . Accessed: February 14th, 2016.

4

The estimated number of connected households by the end of 2014 was 21.7 million, i.e. 62% of the PNBL goal. If we look only at Banda Larga Popular plans (1 Mbps plan for 35.00 BRL), the results are negligible, as we shall see in the next section – “Notes on the results of PNBL specific actions”. For further information on the number of fixed connections at the end of 2014, see COMITÊ GESTOR DA INTERNET NO BRASIL – CGI.br. Research on the use of information and communication technologies in Brazil – ICT. Households 2014 São Paulo: CGI.br, 2015, p. 322324. Available on: . Accessed: February 14th, 2016.

5

Ibid. Chapter 1, Table 1.2 presents data from the International Telecommunications Union showing that in 2014 fixed broadband access densities were higher than 11.5 in countries such as Russia (17.5), China (14.4), Argentina (14.7) and Uruguay (24.6).

80

veridiana Alimonti

TABLE 3.1. Percentage of Brazilian households connected to the Internet with fixed broadband, 2014

Area

Region

Social class

Connected

Fixed broadband

Total

50

33

Urban

54

36

Rural

22

13

Southeast

60

40

Northeast

37

23

South

51

39

North

35

16

Center-West

44

29

A

98

85

B

82

63

C

48

29

DE

14

5

Base: 65.1 million households. Of these, 33.1 million had fixed broadband. Data collected between October/2014 and March/2015 Source: Adapted from CGI.br (2015, Table A4 and A5, p 322 and 323).

The lack of competition in fixed broadband is another critical factor that influences the low density of connections. As of 2016, Brazil had three ma‑ jor groups that together accounted for more than 85% of market.6 They are: América Móvil (Claro/Embratel/Net), Telefônica/Vivo (which bought GVT) and Oi. However, even though these are large groups, that does not imply that their fixed broadband providers are evenly spread among the country’s municipalities. In many cases, there is only one major carrier that operates alongside (or not) other small and medium-sized providers. After analyzing the degree of competitiveness in different municipalities, Oliveira and Figueiredo (2013) concluded that, in 2012, 95% of the country’s cities had concentrated fixed broadband offerings.7 The picture was worse in smaller municipalities, of up to 50,000 inhabitants, where the market was 6 As can be seen in the information consolidated by Teleco Consultoria, based on Anatel data, available on: . Accessed: February 16th, 2016. 7

The researchers conducted the analysis based on the Herfindahl Hirschman Index (HHI), which takes into account not only the number of providers, but also the power of these providers within a given market (OLIVEIRA; FIGUEIREDO, 2013, p. 2930).


81

virtually a monopoly. In the intermediate range between 50,000 and 100,000 inhabitants, almost half of the municipalities fit the same profile. The market was moderately competitive in most cities with populations over 500,000 in‑ habitants. The study also identified an increase in the degree of concentration between 2010 and 2012, the first two years of the PNBL. No doubt, the statistics are much better when we consider mobile broad‑ band access, especially connections via mobile phone. The Ministry of Communications estimates that while access to fixed broadband grew 79% from May 2010 to May 2014, mobile broadband connections rose by about 825% in the same period (MINISTRY OF COMMUNICATIONS, 2015a).8 The euphoria, however, must be tempered in view of the real situation. First, the statistics themselves are questionable, since many mobile voice plans include some type of Internet connection promotion, which was count‑ ed as access to mobile broadband. The problem is that almost 72% of all cell phone plans were prepaid in December 2015, and that includes mobile broad‑ band. In these, data allowances are extremely low, often 10 MB a day, even with 4G speeds.9 The connection ends when the plan runs out. Much of this data is consumed with advertising and updates over which consumers have little control, which limits browsing through prepaid plans. Treating mobile connection as a reasonable substitute for fixed connec‑ tion is a fallacy. The only Brazilian region with more access to the network via mobile devices is the North (IBGE, 2015, p. 34). This is not due to greater tech‑ nological advance in the North, instead, it reiterates the problem of regional inequalities in fixed broadband, as mentioned above.10 Where there is no 4G coverage, the actual speeds rarely reach 1 Mbps, adding to the traditional stability problems in the service. In early 2015, the average speed of mobile connections in Brazil was 2.5 Mbps, well behind the first place, occupied by the United Kingdom (20.4 Mbps), and other Latin American countries, such as Uruguay (5.4 Mbps). When we consider the 8 According to Anatel, the country ended July 2016 with 185 million accesses in mobile broadband, 43 million of which via 4G devices. Available on: . Accessed: September 11th, 2016. 9 For further information on mobile accesses in December 2015, see . Accessed: February 14th, 2016. The four main mobile operators (Claro, Oi, Vivo and Tim) offer prepaid plans with 4G connections, provided that the customer is in the coverage area. 10 Among other devices, cell phones and, to a lesser extent, tablets stand out.

82

veridiana Alimonti

overall speed of broadband connections in the country (fixed and mobile), the average speed rises to 3.4 Mbps (GROSSMANN, 2015a).11 It reached 3.6 in the third quarter of 2015. However, Brazil was still below the world average of 5.1 Mbps, and it ranked 93rd on a list of 145 countries (AKAMAI, 2015, p. 26).

NOTES ON THE RESULTS OF PNBL SPECIFIC ACTIONS The data highlights some of the challenges we face in spreading the access to broadband Internet in Brazil. To complete this scenario, it is worth com‑ menting briefly on the evolution of specific actions conducted by the PNBL: mass broadband plans, connection in rural areas, the job done by Telebras and the REPNBL. Before discussing these specific actions, we must note that the project of maintaining an institutional space for social participation in the PNBL – called Fórum Brasil Conectado – lasted only a few months. It was discontinued at the end of the second term of President Luiz Inácio Lula da Silva. During these months, the forum had a few meetings, but it worked more as a pre‑ sentation space to support government actions than as an actual place for intervention and participation. In Dilma Rousseff’s administration, the relationship with civil society was quite confrontational. Doors were closed to dialogue since the negotiation of the Commitment Terms for Banda Larga Popular plans in 2011. In the sec‑ ond half of the first term, a working group formed by members of the civil society, the Ministry of Communications and the General Secretariat of the Presidency was created. However, this initiative could not present anything valuable in terms of infrastructure-related policies and was terminated shortly after its creation. The President Dilma Rousseff’s willingness to talk changed during her electoral campaign of 2014, when she even participated in an event held by organizations and activists involved with digital inclusion.12 In her second 11 More information on the report from the first quarter of 2015 can be found on: . Accessed: October 6, 2016. Subsequent reports by Akamai, from the second and third quarters of 2015, do not present data on the average speed of mobile connections alone. 12 The event, called Diálogos Conectados, was organized by the Banda Larga é um Direito Seu! campaign. In addition to Rousseff, the then candidate Marina Silva also participated in that initiative.


83

administration, the new Ministry was considering reactivating the Fórum Brasil Conectado or creating a new institutional space for social participation (INSTITUTO TELECOM, 2015).13 Nevertheless, nothing had been done by the beginning of 2016.14 Below there are some considerations on the results of specific actions tak‑ en by the PNBL:

Banda Larga Popular plans These are 1 Mbps plans marketed at 35.00 BRL (or 29.90 BRL wherev‑ er there is ICMS exemption)15 negotiated with the following companies: Telefônica, Oi, CTBC and Sercomtel in 2011. These plans can cost more than 60.00 BRL when combined with telephone service. This bundling was autho‑ rized by the government in the provision of broadband plans, although it is contrary to the Consumer Protection Code (Federal Law 8,078/1990). The results of this program are negligible and not transparent. According to Anatel’s inspection, by the end of the third quarter of 2014 the offer was available in 5,000 Brazilian municipalities (ANATEL, 2014, p. 1). However, the report does not mention the number of Banda Larga Popular subscriptions. The last report of the agency with this information dates back to March 2013. It lists less than 2.6 million contracts (ANATEL, 2013, p. 56). This number decreased in 2014, as pointed out in a survey conducted by a journalist with Anatel and the Ministry of Communications. Of the total fixed broadband

13 The author had the opportunity to confirm this information in a conversation with the Minister of Communications, Ricardo Berzoini, and the Executive Secretary of the Ministry. 14 It is worth mentioning that 2015 was a politically turbulent year, which led to a change of ministers in the Ministry of Communications. Ricardo Berzoini, from the Workers Party (Partido dos Trabalhadores – PT), which initially held the position, became head of the Secretariat (which combined the Secretary-General of the Presidency and the Secretariat of Institutional Relations). André Figueiredo, from the Democratic Labor Party (Partido Democrático Trabalhista – PDT), was appointed Minister in his place. Although the composition of the Telecommunications Department, which deals more directly with broadband policies, has been maintained, we can say that government’s position remains unchanged. 15 The exemption was regulated in only seven Brazilian states (São Paulo, Rio de Janeiro, Espírito Santo, Paraná, Goiás, Pernambuco and Pará). See more details on the topic in Chapter 1.

84

veridiana Alimonti

subscriptions in Brazil in the first half of 2014, only 7.9% were contracted via PNBL, which means 1.84 million connections (BRUNO, 2014).16 Two main factors contribute to the very small number of accesses: (i) operators resisted offering this plan, often hiding it on their websites and call centers;17 and (ii) the setup of the plans itself, with reduced data allowances even in fixed plans and the possibility of operators also requiring subscribers to purchase fixed-line telephone service.

Internet connection in rural areas As already mentioned, the obligation to provide services in rural areas was enforced in 2012, with the auction the 450 MHz frequency band. In this case, the goals were even less ambitious. They provided for the coverage of rural areas in all municipalities by the end of 2015, with connections of 256 Kbps and a 250 MB download allowance. The connection speed should rise gradu‑ ally to 1 Mbps and the data allowance should go up to 500 MB by December 2017. These parameters are incompatible with the currently available network resources and increase the gap between urban and rural areas rather than seeking to reduce it.

Telebras’ Responsibilities One of the measures celebrated by organizations and civil society activ‑ ists when the PNBL was announced was the reactivation of Telebras with the purpose of structuring a national network. A significant part of Telebras’ ac‑ tivities would include offering wholesale links to broadband service providers, with prices set by the government to make this market more competitive. This is because the major carriers selling the service to the end user also negotiate wholesale network capacity to their competitors, charging high prices to make these links available. Providers that contracted with Telebras were required to offer the Banda Larga Popular plan to their customers. 16 In a presentation of the Ministry of Communications in June 2014, the agency informed that in March of that year there were 2.6 million accesses in the mass broadband plan (MINISTRY OF COMMUNICATIONS, 2015a). 17 Ever since providers were required to offer the Banda Larga Popular plan, several surveys were made to monitor compliance. One of them was carried out in 2012 by Idec – Brazilian Institute of Consumer Protection, and several problems were found. The results are available on: . Accessed: April 18th, 2016.


85

Overall, Telebras has shown very little transparency as regards what cit‑ ies and providers were served. Its follow-up plan was also flawed in terms of monitoring these providers in their offering of Banda Larga Popular plans. A report from the Senate on the PNBL (DINIZ, 2014, p. 17) points out that the initial target coverage of 4,278 municipalities by the end of 2014 was not met. By the end of that year, it was expected to serve only 612 cities, 360 of which by direct offering, and 252 and through partners. The report investigated the reasons for this failure and denounced the lack of investment from the federal government as one of them: The main reason for the bad performance is insufficient investment in projects executed by Telebras. The Multi-Year Plan (PPA) 20122015, (…) provides for total investments of 2.9 billion BRL for the PNBL for the period 2012-2013. However, the planning in annual budget laws for the same years provides for investment of only 314.7 million BRL. With the partial freezing (contingenciamento) of fund‑ ing, the value was reduced to 267.9 million BRL. The actual execut‑ ed budget was 214.1 million BRL, i.e. only 7.4% of the expected in the PPA. (DINIZ, 2014, p. 17).

In addition to investments, another setback faced by Telebras was the dif‑ ficulty to carry out the project – set out in Decree 7,175/2010 – to connect the federal public administration. There was opposition from telecom compa‑ nies, which contested the view that the service provided by Telebras could be done without a bidding process. In the end, during the years 2012-2014 much of Telebras spending was concentrated on optical fiber infrastructure deployment for the transmis‑ sion of World Cup matches in 2014. Almost 90 million BRL were used to connect the 12 stadiums, the International Broadcast Center (IBC) and the training centers of all countries (TELEBRAS, 2014). According to Telebras, this network is a legacy for the fulfillment of the PNBL purposes. We need to assess, however, how this infrastructure can be useful to areas near the con‑ nected points and to locations outside the big cities that hosted the World Cup matches.

86

veridiana Alimonti

PNBL Special Taxation Regime (Regime Especial de Tributação do Programa Nacional de Banda Larga – REPNBL) Regulated by Federal Decree 7,921/2013, the initial deadline for opera‑ tors to submit investment projects for telecommunications infrastructure to benefit from these tax incentives was June 30th, 2013. Subsequently, there was an extension for another two years, ending in June 2015. The projects, that receive exemption from the IPI, PIS/PASEP and Cofins tributes, estimate a total investment of 17.7 billion BRL (MINISTRY OF COMMUNICATIONS, 2015b). If the interest of the federal government [in the PNBL] is to stimulate in‑ vestment in telecommunications networks, which would possibly occur to a lesser extent without the tax exemptions, it must be stressed that public funds are not being collected – funds that could be better directed to the task of universalizing, with quality, the access to broadband Internet in the country, including contributions to Telebras, which is subject to budget curtailments.18 Looking at the general list of projects, more than half of the planned funds are concentrated in the Southeast (9 billion BRL).19 The most proposed invest‑ ments are for mobile access (6.4 billion BRL), although only slightly fewer in‑ vestments are destined to transport and optical access (about 5.7 billion BRL). As for the benefited companies, the main proponents are the major carriers, resulting once more in funding that ends up favoring big companies instead of small and medium service providers. Finally, there is the concrete provision for offsetting obligations for the operation of services through these networks, such as the allocation of at least some of them implementing priority public policies. Even the obliga‑ tion originally envisaged to use domestic technology in the network equip‑ ment was reduced in general and lowered to zero for mobile network projects (TELETIME, 2015).

18 The PNBL evaluation report for the Federal Senate highlights that the tax relief estimated by the government at that time was equivalent to 36 times the average annual budget execution of Telebras. It concludes: “in other words, the executive branch destined a fairly high amount of funds to the private sector, while curtailing investments by Telebras”. 19 For the Northeast, the requests amounted to 4 billion BRL, to 2 billion BRL in the South, 1.5 billion BRL in the Center-West and 955 million BRL in the North. The complete list is available on: . Accessed: July 8th, 2015.


87

NEGLECT FOR THE ESSENTIALITY OF BROADBAND AS THE BACKGROUND PROBLEM OF ACCESS POLICIES In 2011, a study done by the Institute of Applied Economic Research (Instituto de Pesquisa Econômica Aplicada – IPEA) stated that one of the short‑ comings of the PNBL was to assume that, by solving the problems of supply at the end of 2014, there would be enough equipment for Internet access and people trained to use it in 35 million households (SOUSA, 2011). The criticism pointed out that, among other measures, it was necessary to invest in education and training for people to use Information and Communication Technologies (ICT). In fact, conceiving digital inclusion as part of a broader process of development involves issues related to digital literacy, accessibility and understanding that including someone digitally means encouraging and enabling mate‑ rial and non-material conditions for each social group to produce its own vision of reality, interpreting, creating, accessing and dissemi‑ nating information capable of qualifying its intervention in the world and enabling people to stand up for their rights. (INTERVOZES, 2012, p 200-201; FERRAZ; LEMOS, 2011)

These assumptions are important so that the expansion of new technol‑ ogies does not deepen inequalities and economic growth actually results in development and social inclusion. This positive effect is not an obvious result of broadband access policies. There must be some articulation between these policies and other actions aimed at empowering people (DIRSI, 2014, p. 92), through integrated communication policies, education, culture, open govern‑ ment, science and innovation, among others. This integrated approach was never really implemented under the PNBL, and this is part of the problem. However, there is a previous question, relat‑ ed to this, but with a specific complexity and which is directly linked to the criticisms already made of specific actions of the program. It is the weakness of the federal government in playing its part as coordinator of the economic activity to overcome inequalities and enforce rights. That’s what we’ll see next.

Legal essentiality regime and regulatory “quick fixes” Even though the PNBL was the result of the federal government under‑ standing that access to broadband Internet is important for economic and

88

veridiana Alimonti

social development, and that the market alone would not be able to properly disseminate such access, the paths taken by the government were very tortu‑ ous, to say the least. Bypassing what Federal Law 9,472 of 1997, the General Telecommunications Law (Lei Geral de Telecomunicações – LGT), says about essential telecommunications services, the federal government chose to re‑ sort to regulatory “quick fixes”. According to art. 65, § 1, of this law, essential telecommunications services are subject to universal service obligations, which “aim to provide any person or institution of public interest with telecommunications services, regardless of their location and socioeconomic status” (art. 79, § 1). Therefore, in accor‑ dance with the regulatory framework established by the LGT, such services should be granted through concessions, with the government having the pre‑ rogative to control their fees and make sure they are affordable. Some degree of continuity is also mandatory, and this applies more strictly to essential ser‑ vices. One of its consequences is the reversion to the federal government, at the end of the concession, of the infrastructure necessary to provide the service. At that time, the government can either exploit that infrastructure or grant a new concession. These are the main features of what the LGT defines as “public regime”, which differs from the “private regime”. In the latter, the approvals for oper‑ ation (outorgas) are granted upon authorization, and the regulation is based on the defense of competition, without obligations related to universaliza‑ tion, reasonable service rates (modicidade tarifária), and revertibility of assets. Currently, fixed telephony is the only telecommunication service provided, in part, under a public regime. Even though broadband20 is increasingly essen‑ tial, including for fixed line telephony, its provision is carried out exclusively under the private regime, in accord with the interest of the companies, which prefer less regulation.21 20 The telecommunications services associated with broadband Internet access are the Multimedia Communication Service (Serviço de Comunicação Multimídia – SCM), which refers to fixed broadband, and the Personal Mobile Service (Serviço Móvel Pessoal – SMP), which relates to mobile broadband. 21 In general, operators defend the private regime because it means less regulation and fewer obligations. Even small and medium providers are opposed to regime change in broadband, for they are afraid to give greater market power to concessionaire companies, which would deepen the concentration. For this reason, we will see that the Banda Larga é um Direito Seu! campaign claims that the public regime focuses on wholesale as a way to promote more competition in retail markets.


89

In spite of that, the federal government has been trying to establish obliga‑ tions related to broadband for several years. The government has been doing this based on its regulatory prerogatives that derive from fixed telephony con‑ cessions. The first significant move in this direction occurred in 2008, when there was a change in the General Plan for Universalization Targets (Plano Geral de Metas para a Universalização – PGMU) for fixed telephony, to replace the installation of Telecommunication Service Stations (Postos de Serviços de Telecomunicação – PST)22 with the deployment of backhaul by December 2010, in all Brazilian municipalities and localities not yet covered. Even though the backhaul23 is part of the network infrastructure for fixed telephony, this ob‑ ligation refers to a demand that is much more related to broadband than to telephony services. In the same year, and in the same context, fixed telephony operators and authorized fixed broadband companies (Telefônica, Brasil Telecom, Oi, CTBC and Sercomtel) signed an amendment to the authorizations with goals for the connection of urban public schools by the end of 2010. In 2011, when the fixed-line concession contracts came up for a third five-year review, and with the PNBL already in force, the same logic was applied to the negotiation of Commitment Terms for the provision of Banda Larga Popular plans. In this way, the federal government uses its prerogatives in the public system (fixed telephony) to negotiate obligations related to another service (broadband), provided on a private basis, taking advantage of the fact that both can use the same transmission infrastructure. Thus, it drains resources from the concession towards investments in other areas. Some problems are the result of this maneuver, the so-called regulatory “quick fix”. The first is the economic and financial imbalance in the fixed telephony concession, which supports spending for the provision of broadband, while only counting as income whatever comes from landline telephone services. One of the consequences is high prices for monthly basic fixed line telephone service (about 40.00 BRL). The second problem is precisely this cross-subsidy between services, which is prohibited by the LGT (art. 103, § 2), through which 22 A PST was a facility open to the public and maintained by the concessionaire, allowing personal customer service and having at least a pay phone, basic dial-up Internet connection and equipment for sending and receiving text, graphics and images. 23 Backhaul is the supporting network infrastructure for broadband connection that links the access networks, all the way up to the end user, to the operator’s backbone. For more detail, see Chapter 2.

90

veridiana Alimonti

a service recognized as essential (fixed telephony) pays for a service rendered exclusively on a private basis (broadband).24 The third is the confusion created between the assets related to the provi‑ sion of each service, especially when the assets related to fixed concession are revertible and represent much of the provision of broadband infrastructure. The backhauls installed under the exchange of goals in 2008 are added to the copper networks where both services are operated. Anatel’s data shows that more than half of all Internet connections in Brazil still have this network as support (GROSSMANN, 2015b).25 Given this confusion, there is a growing movement to alter the LGT so that nothing or very little reverts to the federal government in 2025, when fixed telephony concessions end. If this movement succeeds, investments made as a result of telephony concessions, supported by unjustifiably expen‑ sive subscription fees, will result in assets that will be owned by the compa‑ nies without further consideration of public interest. Thus, there are sectors that criticize the revertability of assets and argue that application of this policy discourages investment, since companies would be investing in assets that will not be theirs in the future.26 However, the economic return is part of the dynamics of revertability. Revertability is considered in the economic and financial balance of the contract and provides for offsetting investments not amortized at the end of the concession, as well as the application of public

24 In technical bulletin 427/PBCPD/PVCPC/CMLCE/PBCP/PVCP/CMLC/SPB/SPV/SCM, from December 5th, 2008, Anatel recognized that more than 80% of the local concession investments were allocated to data service. There are two factors that make it difficult consider these expenditures separately. Initially, the delay of more than ten years for Anatel to develop a cost model for the setting of fees and prices in the exploration of telephony services, as per Federal Decree 4,733 / 2003. Then, the possibility – opened in 2011, therefore before the development of the cost model methodology – of telecom operators integrating the provision of such services in a single company, putting an end to the LGT article that stated that the fixed line operators should be separate companies. 25 It is interesting to note that there is already a certified technology solution in operation. It is called G.fast and combines these copper pairs with fiber optic cables, allowing them to reach high speeds. Thus, this copper infrastructure already installed in Brazil will continue to be relevant at least in the medium and perhaps long-term, ruling out the argument that they are obsolete. 26 As we will see below, one of the proposals under study is the exchange of revertible assets for investments of the operators in their own broadband networks, since the revertability requirement might drive away investment (AQUINO, 2015).


91

funds, from the beginning, to the obligations that imply loss.27 Thus, infra‑ structure funded with public resources and with high fees paid by the people, under the pretext of complying with the economic and financial balance of the concession contract, should be associated with greater guarantees to the State and especially to the society. It should continue to be operated with the public interest and public policies in mind. If the company fails to provide an essential service based on a decision of its own or in the event of bankruptcy, the fate of these assets cannot be at the sole discretion of that company or of the market forces. At the same time, if an essential service is not being provided properly, it is up to the federal gov‑ ernment to ensure quality and continuity, so it should have sufficient powers to do so. As for assets linked to the concession, but not indispensable for providing the service, the law allows them to be sold. But certainly this evaluation must pass the scrutiny of the grantor, and the funds obtained by the company from the sale should be considered in the contract balance to also benefit users. If we consider the best way to ensure continuity of service and public in‑ terests in the operation of these networks in the context of a new concession, it is unreasonable that the billions of BRL in equity arising from current con‑ tracts become exclusive property of the companies.28 To intensify the controversy, this discussion about revertability unfolds in an environment of legal uncertainty. This uncertainty results from the regu‑ latory “quick fix” and affects the instruments that the government uses to im‑ pose fixed broadband obligations on companies, such as amendments to the authorizations and commitment terms with goals that, at first, seem strange to the private regime configuration. Thus, the dynamics of mandatory peri‑ odic review of these goals and the clear ability to control the rates and fees, among others, do not include broadband. Loose and weak regulatory instru‑ ments also result in loose and weak commitments. The correct would be that

27 This last element has been affected by the historical diversion of sector funds, especially Fust, to other purposes by the Ministry of Finance, a practice that should be corrected so that the system can function properly. 28 Although part of the assets of the concessionaire for the duration of the concession agreement, the revertible assets are of public interest and, according to the LGT, their ownership is automatically given to the federal government at the end of the contracts. The objective of revertability is to assure continuity of service and a new bidding process focused on the public interest.

92

veridiana Alimonti

the telecommunications services that support broadband, especially the fixed service (SCM), were also provided in the public regime. Despite the approval in April 2014 of the Brazilian Civil Rights Framework for the Internet (Marco Civil da Internet, Federal Law 12,965), which recognized the essential nature of Internet access for all citizens and established that the Internet use in Brazil aims to ensure Internet access rights to all, the government seems to insist on a tortuous path.29 An alternative is also being discussed. At the end of 2015, the Ministry of Communications opened a public consultation30 the result of which will be crucial for the definition of future policies for broadband, as well as for the regulation of telecommunications in Brazil. The direction of the Broadband for All program, announced by Dilma Rousseff during her reelection campaign, depends directly on the conclusions of this process.

Broadband for All program and regulatory model under debate According to the Ministry of Communications, the Broadband for All pro‑ gram – still on hold and without official release as of March 2016 – aims to reach 90% of the municipalities and 45% of urban households with fiber optic networks, with an average speed of 25 Mbps, by 2018. Projects already underway by Telebras are included in the program, such as the submarine cable between Brazil and the European continent and the launching of a geo‑ stationary satellite.31 The program will be funded by a combination of three main mechanisms: (i) direct funding from the federal government for the actions of Telebras; (ii) reverse auctions for optical transport and access networks, where the 29 In the proposed PGMU IV – presented by Anatel for public consultation in the context of the last five-year review of the fixed telephony concession contracts – we can find the obligation to deploy fiber optic backhaul in municipalities and localities that do not have such infrastructure. This review process should have ended at the end of 2015. However, because of the public debate on the regulatory framework for the telecommunications sector, to be addressed in the next section, the completion of this review was postponed to April 2016. When it comes to the item of the proposal discussed here, it is worth mentioning that some civil society organizations and activists following the issue do not object the inclusion of backhaul goals in fixed telephony concessions, especially because of the delay or even the denial from the government to establish the provision of broadband also in the public regime. 30 More information about the public consultation on: . Accessed: February 9th, 2016. 31 Under the program, it will be up to Telebras to deploy the private network of the federal administration, in addition to continue to supply network capacity in wholesale to other broadband service providers.


93

winner will be the company that needs the lowest public subsidy for invest‑ ments and to meet the required goals; and (iii) creation of credit lines from the National Bank for Economic and Social Development (Banco Nacional de Desenvolvimento Econômico e Social – BNDES) to finance the optical transport and access networks, including those of small and medium providers. For reverse auctions, the federal government expects to finance the public subsidy with tax credits related to the Telecommunication Inspection Fund (Fundo de Fiscalização das Telecomunicações – Fistel). At this point, it is worth noting that it is Fistel and not the Fund for Universalizing Telecommunication Services (Fundo de Universalização de Serviços de Telecomunicação – Fust) that the government intends to use. This happens be‑ cause of a limitation found in the Fust Law (Federal Law 9,998/2000). As shown in its name, this fund is intended to cover investments for universal services and therefore can only be applied to telecommunications services in the public regime, i.e. in fixed telephony. There is a bill to change this (Draft Law 1,481 / 2007), which would not be necessary if broadband services were also provided in the public regime. That would be the most suitable solution. In addition to the legal definition, both Fust and Fistel suffer from an earlier problem – the diversion of the revenue they generate in order to meet federal primary surplus goals. In the case of Fistel, revenues in recent years were about 5 billion BRL annually, rising whenever another radio frequency is auctioned. Fust, in turn, collected nearly 1.8 billion BRL in 2014. If we consider figures ever since they were created, in 2015 Fistel would have had 58.2 billion BRL and Fust over 20 billion BRL (TELEBRASIL, 2015, p. 151).32 However, the effective application of these funds in the sector is neg‑ ligible, which is another demonstration of the lack of priority given to tele‑ communications by the federal government (see Chapter 8 for more details). Still considering the Broadband for All program, there were no provisions for changing the delivery regime of telecommunications services related to broad‑ band to also include its provision under the public regime. At first, they would continue to be provided under private regime, through permits/authorizations that would have a slightly different type of obligation for goals and investments. The approval of the program through a specific law was also considered. Under pressure to put that policy into practice and the approaching end of fixed-line concession contracts, with all their imbroglios, in November 2015 32 Chapter 8 provides details on these funds and presents a full table of tax collection over the years.

94

veridiana Alimonti

the Ministry of Communications launched a public consultation to review the current delivery model of telecommunication services. The consultation did not have any a priori model. However, its “basic text” contains some biases, such as restrictions on the use of concession contracts and in particular the maintenance of asset revertability. The deadline for contributions ended in January 2016. The government’s goal was to develop a bill to be sent to the Chamber of Deputies (lower house of Congress) as part of the discussions on the reform of the General Telecommunications Law. These discussions were held throughout 2015 in a special subcommittee of the Committee on Science, Technology, Communication and Informatics (Comissão de Ciência e Tecnologia, Comunicação e Informática (CCTCI) of the Chamber of Deputies.33 The initiative of the Ministry clearly focused on creating a new regulatory framework for the sector, with emphasis on broadband. One of the questions posed is whether it still makes sense to differentiate the public regime from the private regime in the provision of telecommunications services. In fact, the opening of this public consultation was the formalization of a debate that had already occurred for some time behind the scenes but became more heated with the approach of the five-year review of the con‑ cession contracts. In the first half of 2015, three proposals were discussed internally and resonated in the government (especially the Ministry of Communications and Anatel):34 1. One advocated the end of the revertability of the assets and the trans‑ formation of existing concessions into authorizations by decree or bill. The value of the revertible assets would be calculated, no matter how difficult that would be, and transformed into investment goals for broadband. 33 It is the Special Subcommittee on Mobile Services and Pay TV. More details are available on: . Accessed: February 9th, 2016. 34 The listed proposals are based on news about an internal document from the Ministry of Communications. See Luís Osvaldo Grossmann, “Minicom rascunha Lei de Universalização da Banda Larga em regime público”, Convergência Digital, June 26th, 2015. Available on: . Accessed: July 9th, 2015. Also see Miriam Aquino, “Estudo do Minicom justifica sua opção por banda larga pública”, Tele.síntese, July 10th, 2015. Available on: . Accessed: July 12th, 2015.


95

2. Another proposal added to the end of the revertability an amendment to the LGT to eliminate the difference between public and private re‑ gime. The provision of all telecommunications services would fall un‑ der a single regime, more similar to what is now the private regime. 3. The third approach to the public regime advocated changing the LGT to allow successive extensions of telephony concessions, or at least one further extension, which would include broadband in the object of ex‑ tended contracts. The revertability of assets would be maintained and rates could be controlled only where the market was not competitive. According to the technical area of the Ministry of Communications, this would be the proposal that best solves the complex situation pre‑ sented in the previous item. In fact, the third proposal does not face the legal problem it might face if the federal government decided to dispose of billions of BRLs of equity in revertible assets, exchanging it for investment in the telecommunications companies’ private networks. However, this approach is subject to another question: the change of the object of fixed telephony concessions without rebidding – which was already addressed by the Court of Audit (POSSEBON, 2015). The path taken by the government to conduct this debate and come up with a definition was therefore launching the public consultation, which should result in a draft law and a set of guidelines for the establishment of public policies in the field of telecommunications. Whether the changes will be for the worse or for the better, it is still early to say. In this context, the last section of this chapter presents alternatives and positions from the civil society perspective, in particular what has already been developed under the Banda Larga é um Direito Seu! campaign.

HOW TO PROCEED? Since the PNBL was launched, a group of organizations and civil society rep‑ resentatives have been monitoring the issue and seeking to influence the out‑ come. The Banda Larga é um Direito Seu! campaign35 brings together organiza‑ tions like Intervozes, Idec, Proteste, Centro de Estudos da Mídia Alternativa Barão 35 More information about the campaign and its organizations can be found on: . Accessed: February 9th, 2016.

96

veridiana Alimonti

de Itararé, Fórum Nacional pela Democratização da Comunicação (FNDC), Clube de Engenharia, Coletivo Digital and Capítulo Brasil da União Latina da Economia Política da Informação, da Comunicação e da Cultura (Ulepicc). In 2012, the campaign presented to the Ministry of Communications and Anatel a proposal that part of the provision of fixed broadband service be un‑ der the public regime. The proposal suggests a layered approach, separating what was called the “network layer” (mainly the operation of large trunks – backbone and backhaul) from the so-called “service layer” (the offer to the end user).36 Today, SCM is the telecommunications service that supports access to the Internet in both layers. In this proposal, the network layer should match another service provided under the public regime. This eminently wholesale service should be operated as a public concession; the concessionaires would then observe the universalization targets related to the deployment of back‑ bones and backhaul. The wholesale of network capacity by these concession‑ aires would be regulated by a reference price, ensuring an appropriate value for the link. In turn, the contracting providers in the last mile would remain operating on a private basis, but they would be required to offer a basic plan to their customers. The characteristics of the plan (including price) would be defined by the government to ensure adequate access to the Internet, with speeds compatible to the currently available network resources. If the availability of the link in wholesale at the reference price did not stimulate the provision of service at the retail level due to a lack of providers willing or able to do so, one of the alternatives would be to also include invest‑ ment obligations in the concession of the “network layer”. Other alternatives would be the construction of access networks by Telebras, directly providing the service to the end user or granting the use of these networks to local providers. There could also be reverse auctions with the provision of public funds to encourage investment in these networks by small and medium providers, in accordance with goals for service expansion and other obligations contained in the public notice of such auctions. With the same objective, we propose a functional separation between the concessionaire of the network layer and the providers of retail broadband ser‑ vice. In other words, they should come from different economic groups or dif‑ ferent companies within the same group. In case of the latter, the conditions applied by the concessionaire to the partner within its group should be the same 36 The proposal in its entirety can be accessed on: . Accessed: December 19th, 2015.


97

as those offered to other providers. The reference price set by the grantor must always be respected. That would contribute to a more competitive environment in the last mile, which will favor the sustainability of the providers, not only commercial providers, but also community and local government providers. In addition to setting up a basic plan to be offered, public policies and obligations of the concessionaire should include collective access goals, such as the spread of Wi-Fi hot spots and public access centers, training and collab‑ orative production of digital content. The investments and the returns of the concessionaire would match the economic and financial balance established in the concession contract, using Fust revenues for the portion of investments not yet amortized so as to assure efficient operation of broadband service. In this case, there would be no ob‑ stacle to the use of Fust, since the service would be provided under the public regime. Finally, it is essential to ensure institutional spaces for effective social participation in monitoring the development and implementation of infra‑ structure and digital inclusion policies. For the transition of regimes, the campaign proposes the establishment of a General Concessions Plan (Plano Geral de Outorgas – PGO). This plan would focus on the “network layer” and consider the distribution of current fixed line operators without excluding the entry of new operators. Broadband networks defined in this plan would then be included in the list of assets af‑ fected by the grant. Companies would be compensated by infrastructure unre‑ lated to fixed universalization obligations. The ratio of cross-subsidy between telephony and fixed broadband would also be discounted. Finally, according to art. 18, I of the LGT, the provision of that service under the public regime system could be achieved by decree, with no need for a new law. The proposal was made in accordance with the current regulatory model and therefore takes into account the distinction between the public and private re‑ gimes for the provision of telecommunications services. The organizations that make up the campaign do not view this distinction as fundamental. In fact, many of them believe that the private regime as outlined ultimately weakens the char‑ acterization of telecommunications services as a public service/utility (as estab‑ lished in art. 21 inc. XI of the Federal Constitution). Thus, the possibility of estab‑ lishing a new regulatory framework without separate regimes seems interesting, as long as the current elements defining the public regime continue to be applied to essential telecommunications services, as is the case of broadband.

98

veridiana Alimonti

Any solution that results in eliminating these elements will be incompat‑ ible with the essential part of telecommunications services and will conflict with our constitutional order.37 The Banda Larga é um Direito Seu! campaign takes the position that it is not possible to provide widespread access to broadband without the gov‑ ernment having the proper regulatory instruments. Moreover, access is only part of the task – digital inclusion policies are needed to promote training, stimulate innovation and create applications in order to promote technology ownership for empowerment and development. As relevant as the result of the debate on the regulatory model for the sector is its process, which should be carried out openly and with broad participation, given that it relates to a strategic infrastructure and to the future of a service intrinsically linked to the guarantee of fundamental rights in the country.

REFERENCES AGÊNCIA NACIONAL DE TELECOMUNICAÇÕES (Anatel). Termos de Compromisso aderentes aos objetivos do PNBL firmados entre MC, Anatel, Oi, Telefônica, Sercomtel e CTBC – Executive Report. Brasília, March 14th 2013. Available on: . Accessed: July 3rd, 2015. ______. Programa Nacional de Banda Larga – 3rd quarter of 2014. Brasilia, November 19th, 2014. Available on: . Accessed: July 3rd, 2015. AKAMAI. Akamai’s State of the Internet 3rd Quarter 2015 v. 8, no. 1, 2015. Available at: . Accessed: October 6, 2016. AQUINO, M. Bechara: to Estado deve acabar com a concessão de telefonia e reverter a posse. Telesíntese, July 21st, 2015. Available on: . Accessed: August 1st, 2015. 37 The full Intervozes – Coletivo Brasil de Comunicação Social – contribution to the public consultation on the review telecommunications services provision model is available on: . Accessed on: February 9th, 2016.


99

BRUNO, L. broadband program draws to a close criticized by authorities and operators. Reuters Brazil 30 September 2014. Available on: . Accessed: July 3rd, 2015. COMITÊ GESTOR DA INTERNET NO BRASIL (CGI.br). Survey on the use of information and communication technologies in Brazilian Households – ICT Households 2014. Sao Paulo: CGI.br, 2015. Available on: . Accessed: February 14th, 2016. COMITÊ GESTOR DO PROGRAMA DE INCLUSÃO DIGITAL (CGPID). Brasil Conectado: Programa Nacional de Banda Larga. Brasília, 2010. DIÁLOGO REGIONAL SOBRE SOCIEDAD DE LA INFORMACIÓN (DIRSI). The Internet and Poverty: Opening the Black Box. In: Galperin, H.; MARISCAL, J.; BARRANTES, R. (eds.). IDRC / CRD, July. 2014. Available on: . Accessed: July 8th, 2015. DINIZ, A. Relatório de Avaliação do Programa Nacional de Banda Larga (PNBL). Comissão de Ciência, Tecnologia, Inovação, Comunicação e Informática do Senado Federal. Brasília. December 2nd, 2014. Available on: . Accessed: July 4th, 2015. FERRAZ, J.; LEMOS, R. Pontos de Cultura e Lan Houses: estruturas para inovação na base da pirâmide social. Rio de Janeiro: FGV Law Rio, 2011. Available on: . Accessed: July 7th, 2015. ______. Rede de cobre ainda concentra 53% da banda larga no Brasil. Convergência Digital, July 1st. 2015b. Available on: . Accessed: July 9th, 2015. INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA (IBGE). Pesquisa Nacional por Amostra de Domicílios: acesso à Internet e à televisão e posse de telefone móvel celular para uso pessoal. Rio de Janeiro, 2015. Available on: . Accessed: July 3rd, 2015. INSTITUTO TELECOM. A volta do Fórum Brasil Conectado. March 16th, 2015. Available on: . Accessed: August 29th, 2015. INTERVOZES – Coletivo Brasil de Comunicação Social. Caminhos para a Universalização da Banda Larga: experiências internacionais de desafios brasileiros. PEREIRA, Sivaldo; BIONDI, Antonio (orgs.). São Paulo: Intervozes/

100 veridiana Alimonti

CGI.br,2012. Available on: . Accessed: July 1st, 2015. MINISTÉRIO DAS COMUNICAÇÕES. Programa Nacional de Banda Larga – Balanço 20102014, May 2015a. Available on: ;. Accessed: July 29th, 2015. ______. REPNBL garante R$ 10 bilhões em investimentos. July 3rd, 2015b. Available on: . Accessed: July 8th, 2015. OLIVEIRA, J. M.; FIGUEIREDO, C. O. Análise dos Determinantes da Demanda por Conexões de Banda Larga Fixa no Brasil. In: Radar: tecnologia, produção e comércio exterior, Brasília, Ipea, n. 30, Dec. 2013, p. 2533. Available on: . Accessed: July 2nd, 2015. POSSEBON, S. Para ministro do TCU, mudança no objeto do contrato de concessão exige nova licitação. Teletime, 17 jun. 2015. Available on: . Accessed: January 16th, 2016. SOUSA, R. A. F. Futuros desafios para o Programa Nacional de Banda Larga. Radar: tecnologia, produção e comércio exterior, Brasília, Ipea, n.15, ago. 2011, p. 9-18. Available on: . Accessed: July 1st, 2015. TELEBRAS. Telebras fecha em R$ 89,9 milhões os investimentos na rede da Copa do Mundo, 11 set. 2014 September 2014. Available on: . Accessed: July 4th, 2015. TELEBRASIL. O Desempenho do Setor de Telecomunicações no Brasil: séries Temporais 9M15. Available on: . Accessed: February 14th, 2016. TELETIME. Minicom reduz a zero exigência de tecnologia nacional para projetos de redes móveis, Jan 14th, 2015. Available on: . Accessed: October 6, 2016.

|4|

GENERAL PLAN OF COMPETITION GOALS ABRAÃO BALBINO E SILVA [ National Telecommunications Agency – Anatel ]

T

INTRODUCTION

his chapter presents the General Plan of Competition Goals (Plano Geral de Metas de Competição – PGMC) proposed by Anatel in 2012. We will present its results, characteristics, theoretical basis, and the criticism from the market, especially with regard to the promotion of compe‑ tition in broadband services. It is clear that today’s society has a strong demand for telecommunications services. Access to communication and information systems has become a basic requirement for virtually all sectors of the economy. With the emergence of broadband networks, each year new products, ser‑ vices, and digital content are released. New demands are also identified, so the importance of networks and systems is increasing. On the other hand, telecommunications demand intensive infrastruc‑ ture, with sunk costs, economies of scale, externalities, and assets that can be seen as public property. In addition to that, the sector has to deal with breakthrough technologies and the dynamism associated with these types of services. Therefore, this environment needs intense regulation to ensure re‑ alization of gains from economies of scale inherent in the industry, but also innovation and the emergence of new players. Over the years, the telecommunications industry has undergone many changes, so many of its monopolistic characteristics – such as the monopoly of the access network and the monopoly of call termination – no longer exist. New wireless technologies using a wide range of radio frequencies enable the existence of multiple service platforms that offer the same benefit to the end user. The Internet and applications that make calls allow a single user to be reached in several ways.

102

abraão balbino e silva

This dynamic allowed the emergence of competition in some segments, such as mobile telephony, however, it could not remove all the barriers to new market entrants. On the contrary, there are still some issues related to capital intensity and economies of scale and scope, thus making the telecommunica‑ tions market very concentrated. In this situation, on the one hand we see the need for innovation, caused by the demand for new services and, on the other hand, we can see that the sector has serious concentration issues. That is the environment in which the promotion of competition is increasingly important. Economic theory clearly establishes the importance and the benefits of competition to the markets. The booklet of the Administrative Council for Economic Defense (Conselho Administrativo de Defesa Econômica – Cade), highlights that: Competition is essential in the context of a market economy: it en‑ ables an increase in the variety and quality of products, while at the same time it helps reduce their prices. Competition is the key factor allowing prices to reflect a balanced relationship between supply and demand. In order to obtain the benefits derived from competition, companies must invest in technology and conduct market research to under‑ stand and meet the expectations of their customers. Competition is a tool that can be used to benefit the population. Since the citizens are the end consumers of products and services, it is the citizens that experience the improvements resulting from enhanced competition. In addition to providing benefits to consumers, the dispute between companies in a competitive environment helps the Brazilian econo‑ my in a foreign market structure. (MARQUES, 2004, p. 359)

The PGMC is the main regulatory tool in Brazil to promote competition in the telecommunications industry. The PGMC includes a set of measures established by Anatel that aim to cover and act in different dimensions to pro‑ mote competition while not discouraging investment in the telecom market.

Broadband supply chain The provision of broadband services has a well-known supply chain. A ser‑ vice provider that seeks to work in this industry must have not only an access network but also metropolitan infrastructure capable of linking to this access

general plan of competition goals

103

network. It must also have transport infrastructure capable of connecting this network to the main nodes of connectivity. This infrastructure makes up the so-called Internet ecosystem, which comprises not only small and large Internet service providers, but also con‑ tent and application providers, infrastructure companies, data centers etc. In a simplified fashion, the supply chain for broadband could be defined as follows:

FIGURE 4.1. Broadband supply chain National/ international connectivity

Long-distance transport

Metropolitan transportation

Local access network

Source: Anatel

The local access network comprises the equipment and the infrastructure designed to connect end users, the so-called last mile. Historically, traditional telephone networks were the only option for local access networks. However, a number of new technologies – such as coaxial cable networks (used for pay TV), wireless networks (3G mobile networks, 4G, Wi-Fi, radio etc.) and, more recently, fiber optic cables – have brought forth new alternatives for providing better service. Metropolitan transportation comprises all links and infrastructure, includ‑ ing switches and routers, used to carry the signals and data trafficked from the last mile. It usually corresponds to the connections of various wire centers or base stations to network operation centers or stations of various providers. With the evolution of technology, metropolitan transport is basically com‑ posed of high capacity links, usually radio or fiber optics. Long-distance transport basically plays the same role as metropolitan trans‑ port, however, it uses high-capacity backbones that connect the various met‑ ropolitan centers, as well as radio or satellite links. National and international connectivity corresponds to interconnection with other providers that understand this universe, either for exchanging traffic or for traffic to be delivered to other networks.

104


Since there is a wide range of players, this link in the chain is highly com‑ plex, with groups operating at the international level with infrastructure and networks interconnected worldwide, as well as small or large local groups. Moreover, we could say that there is, in addition to all of these links in the chain, a particular link that corresponds to the passive infrastructure support‑ ing this infrastructure. Here we can mention the towers, poles, ducts, con‑ duits, trenches, and all civilian infrastructure that supports the construction of networks. This passive infrastructure is a key link. Understanding it from the point of view of competition promotion is also important.

The ladder of investment Professor Martin Cave, among others, has developed a regulatory frame‑ work for this “ladder” of investment in infrastructure (CAVE, 2006, p. 223237). In this model, it is assumed that the investment of new entrants is made step by step, until they reach the same level as traditional players. However, for these players to invest in their infrastructure, a range of com‑ plementary products is necessary to build a customer base and enable such companies to offer their services to end users. In other words, this model says that for the promotion of competition there must be some conditions to allow new entrants to have access to parts of the infrastructure that could not ini‑ tially be built by these entrants. Thus, new companies should have access to infrastructure that is not theirs. Over time and with the development of their business, conditions arise for that investment in the other steps and value chain infrastructure. This model is widely used throughout the world. The European Union uses it as a guideline for its member countries.1

1

The European Commission and its various directives apply the model. Market analysis and recommendations of the European Commission to be followed by Member States in the regulation of telecommunications markets are available on: . Accessed: January 14th, 2016.


105

FIGURE 4.2. The ladder of investment

Ladder of investment

Access network

own network

unbundling Bitstream access resale

Transport network

own infrastructure

passive infrastructure

Dark fibers dedicated lines

Source: Anatel

For broadband services, this model would apply to the possibility of new providers using elements from the broadband value chain that belong to a larger operator, until these new entrants were finally able to build their own infrastructures. In a simplified fashion, consider a company that initially leas‑ es all links in the chain (access network, metropolitan transport, long-distance transport and national/international connectivity). Thus, such a company leases the entire telecommunications infrastructure of third providers, acting like a dealer. Over time, as its business gains strength, the company in ques‑ tion can start building its own network, investing in areas with higher value added. Finally, investments in transport network and direct connectivity with leading providers become the boundaries of this ladder of investment. Therefore, the strategy for promotion of competition in broadband seems to be all about reducing barriers to the entry of new companies. This could be done by giving them access to assets they do not have or cannot build, at least not at the very start of their operations. However, some questions come to the fore: • What companies should be required to share their networks with new players?

106


• In what areas and under what conditions should this sharing be done? • To what extent would this model promote inefficient competitors in‑ stead of promoting competition?

Market dominance In general, strategies to promote competition seek to balance the playing field between various existing players in the market, creating the so-called lev‑ el playing field, i.e. giving competitive conditions to all providers. The ladder of investment model emphasizes the reduction of barriers to entry of new providers. However, we must think carefully about its implementation. As already presented, the telecommunications industry is very capital in‑ tensive, with large economies of scale and scope. The simple requirements of network openness and shared infrastructure, if applied indiscriminately, could discourage investment and foster the existence of “hitchhiking” pro‑ viders that would have no real incentive to invest or innovate. Furthermore, economies of scale and the scope of the telecommunications industry can‑ not be considered harmful to the competition itself. These gains can improve efficiency so that end users can get further benefits. Thus, pro-competitive measures should only be applied in particular circumstances. It is necessary to promote competition and innovation, but without discouraging efficient players in their investments and their search for efficiency gains from vertical integration, economies of scale, and scope. In order to understand this relationship we use the concept of “market dominance”. The concept of dominance in the market is a measure of the strength of a brand, product, service or company in relation to its competitors in a “relevant market”.2 A relevant market is in a geographic location where products and companies compete in a certain business. Thus, to define mar‑ ket dominance we must see to what extent a product, brand or company has any control of a product category’s supply chain in a given geographical area. 2

This definition is commonly used in the defense of competition laws and manuals in several countries. This term is also often mistaken for the term “market power”. Both expressions are equivalent, however, the term “market power” has a clearer economic meaning. Massimo Motta, in his Competition Policy: Theory and Practice (2004), defines market power as the ability of a company to increase its prices above a competitive level in order to increase its profits. The same author argues that the concept of market dominance must be considered stronger than market power. It should be used in situations where market power is so strong that it allows an almost monopolistic behavior.


107

From this concept we can understand that if any dominant company be‑ comes so powerful that the competition in that market is impaired, there is in‑ deed the need to adopt measures to mitigate or avoid risks and damage. Thus, the concept of dominance serves as a criterion for the adoption of pro-com‑ petitive measures. Markets that do not have a high risk of overreaching power dominance probably have reasonable competition characteristics and do not require additional measures. To determine whether market dominance exists, there is a test common‑ ly used, especially by the European Commission,3 the so-called three criteria test. This test consists of analyzing a particular market on the basis of three different criteria. Only when the three of them are met should this market be subject to pro-competition regulatory measures. The three criteria are: • The existence of high and non-transitory barriers to the entry of a com‑ petitor – these barriers may be structural, such as technically unfeasi‑ ble network or spectrum limit implementation – or they may be more recurrent, with regulatory, legal or economic issues (such as questions of scale and scope). • The market structure does not tend to have effective competition, not even in the long term. In other words, the current players in this market do not tend to have a satisfactory level of competition with each other. Some criteria and indicators are presented below. • The enforcement of traditional laws and rules to protect competition does not fix this market failure. In this case, it might be impossible to detect concrete faults because of information asymmetries. There might also be situations where an ex post reaction cannot prevent significant dam‑ age to market, like companies that have not even entered the market due to the failure. This type of modeling, therefore, analyzes the various markets in the tele‑ communications industry, with their respective value chains. The analysis should verify the existence of barriers to entry and identify their causes, whether 3 The European Commission systematically adopts this criterion for the definition of relevant markets subject to regulation ex ante since 2003. The latest guideline was published in October 2014 and recommended the use of these criteria for identifying relevant markets. The document is available on: . Accessed: January 14th, 2016.

108


or not they are transient and if they can be dealt with according to traditional laws. After this process, it is possible to identify the providers with dominant power in each relevant market and take action to encourage competition. Usually companies that have market dominance have some specific fea‑ tures, which serve as criteria for analysis after the relevant market is deter‑ mined. They have the following characteristics: • High market share, typically more than 20%; • Control over infrastructure that is difficult or impossible to duplicate, the so-called essential facility (critical infrastructure); • Large economies of scale and scope in the supply of products; • A high degree of vertical integration, controlling supplies that other players need; • Easier access to credit than other players. Therefore, the market dominance model serves as a tool for implementing pro-competition policies. If the relevant markets and companies with domi‑ nance are identified, we can plan measures to allow the entry of new players and mitigate the risks of excess market power.

PGMC The PGMC, approved by Resolution 600/2012,4 is a set of measures ad‑ opted by Anatel in order to promote competition in the telecommunications market. This regulation seeks to put competition promotion theory above into practice. Despite being a regulation from 2012, the principles of the PGMC date back to the early days of industry regulations in Brazil.5 The logic of the PGMC follows the market analysis recommended by the triple test model, with the objective of: (i) identifying markets with excess power that need pro-competitive measures; (ii) identifying companies that have dominance in these markets – the so-called companies with significant 4

The structure of PGMC, acts of Anatel and related documents are available on: . Accessed: January 14th, 2016.

5

Several specific regulations aimed at promoting competition were published by Anatel over the years, namely: Regulamentos de Exploração de Linha Dedicada (Resolutions 402/2005, 437/2006 and 590/2012…), Regulamentos de Remuneração de Redes (Resolutions 250/2000, 319/2002, 438/2006, 576/2011).


109

market power (SMP); (iii) defining regulatory measures to be applied to these SMPs – the so-called “asymmetrical regulatory measures.”

The relevant markets identified and groups with SMP After studying all the value chain of the telecommunications industry, with special attention to the broadband supply chain presented above, the PGMC established the existence of five relevant markets subject to the appli‑ cation of asymmetric measures. The markets identified are:6 • Fixed access network for data transmission via copper networks and coaxial cable at rates lower than 10 Mbps (local access market); • Fixed transport network for data transmission at rates lower than 34 Mbps (market network of local transport and long distance); • National roaming; • Termination on mobile networks; and • Passive infrastructure. The local access market is the provision of access network to providers interested in broadband and fixed telephony services. The products offered in this market are basically full unbundling, which is the breakdown of the pair of copper wires running from a switching center to the end user’s address, and the bitstream, which is the logical breakdown of the access network, al‑ lowing a third party provider to have logical access to the end user’s address. The market for local and long-distance transport network is for providing dedicated links, the so-called “dedicated line industrial exploitation circuits” [exploração industrial de linha dedicada – EILD] and backhaul for various dis‑ tances, as well as for offering IP connectivity to other data networks (Class V Interconnection and Interlinking). The markets for domestic roaming and termination on mobile networks are oriented to the mobile value chain. In the first case, it is the SMP for supplying data and voice roaming service to areas where the buyer has no

6 The technical study by Anatel for each of the relevant markets, the analyzed and nonselected markets, and the application of the criteria defined by the PGMC is available on: . Accessed: December 14th, 2015.

110


coverage. The second case refers to the interconnection of mobile networks, one of the typical markets for the enforcement of regulatory asymmetries. The passive infrastructure market includes all the basic inputs for net‑ work construction, more specifically towers, pipes, ducts and ditches. Figure 4.3 summarizes the main economic groups identified with SMP in Brazil.7

FIGURE 4.3. Main economic groups identified with SMP in Brazil Telefonica / Vivo Group • Local fixed access • Local and LD transport • Passive infra • Roaming • Mobile interconnection

Telecom / Italia Group • Passive infra • Roaming • Mobile interconnection

Oi Group • Local fixed access • Local and LD transport • Passive infra • Roaming • Mobile interconnection

Algar Telecom Group • Local fixed access • Local transport • Passive infra

America Movil Group • • • • •

Local fixed access LD transport Passive infra Roaming Mobile interconnection

Copel Group • Local fixed access • Local and LD transport • Passive infra

Source: Anatel.

Proposed asymmetries and expected results The PGMC established a set of asymmetric regulatory measures imposed on the economic groups that have SMP. The measures imposed can be de‑ scribed as transparency, equality and price control measures. In other words, Anatel determined that in markets where SMP is identified, the common in‑ puts in these markets, the so-called “wholesale products”,8 should be offered

7

The acts that bring the designation of company groups having SMP are available on: . Accessed: January 14th, 2016.

8

As defined by the PGMC “Products in the wholesale market: backhaul, industrial exploitation of dedicated lines (EILD), interconnections and interlinks of telecommunications networks, breakdown of logical channels (bitstream), full unbundling of the local loop, and infrastructure composed of ducts, conduits, poles and towers, under the regulations.”


111

in a transparent and clear manner, without discrimination among buyers and under commercial conditions controlled by the regulator. To ensure the first two measures, Anatel ordered the creation of a Supervisory Entity for Wholesale Offers (Entidade Supervisora de Ofertas de Atacado – ESOA), an independent body created specifically for this purpose. It should be funded by groups with SMP, for monitoring product offerings in the wholesale market and the organization and monitoring of order queues. That body shall operate the so-called Wholesale Negotiation System (Sistema de Negociação de Ofertas de Atacado – SNOA). This system must include all the offers and negotiations from all companies with SMP, in a centralized fash‑ ion, to ensure equality, transparency and non-discrimination in the wholesale products supply chain. To ensure price control, Anatel determined that for each product there must be a Reference Offer for Wholesale Products (Oferta de Referência de Produto de Atacado – ORPA) approved by the agency. All negotiations should be based on that reference offer, which must pass tests to ensure that the price and other commercial terms enable competition.9 These measures are expect‑ ed to encourage competition in areas where SMP is identified. Thus, we expect to mitigate the risks of market power abuse by dominant companies and reduce barriers to the entry of new players.

WHOLESALE NEGOTIATION SYSTEM (SNOA) The SNOA is one of the main tools for promoting the competition sought by the PGMC. It is a unique electronic system maintained and managed by telecommunications service providers, under the coordination of Anatel. In this system, sales of wholesale products defined by the PGMC are negotiated. The SNOA contains the records of all switching centers for SMP providers (includ‑ ing the locations served by these centers), from all points of interconnection (POI), traffic exchange points (PTT) and all towers available for marketing by these providers. This means that – on a single website – any provider seeking Available on: . Accessed: April 18th, 2016. 9 The PGMC says that the price of the Reference Offers must have a characteristic called “replicability”. Therefore, Anatel must ensure that, whenever a Reference Offer is approved, each available product is “replicable”, so that the applicant company is able to compete in the retail market. However, there are no published regulations that define exactly what criteria should be used to measure this characteristic.

112


access to a telecommunications facility can see the infrastructure available, the commercial terms of access to these products, as well as the deadlines and other conditions of service set out in an reference approved by Anatel. Figure 4.4 shows the conceptual structure of the system as intended by the PGMC.

FIGURE 4.4. Conceptual framework of SNOA and its ecosystem Operators Requesting Wholesale Products

Standardized Interface for the GIESB


Centralized Access to BDAs

ESOA (Supervising Entity)

SNOA (Siystem for Negotiating Wholesale Offers)



BDA

BDA Operator with SMP

Operator without SMP Source: Anatel.

The PGMC says that each provider with SMP must maintain a wholesale database (base de dados de atacado – BDA), integrated with the SNOA, contain‑ ing information on all negotiations between that provider with SMP providers and requesters. The ecosystem, with information about all the negotiations between providers with SMP and their respective requesters, is operated by a supervisory authority, and the data and transactions are available online to be reviewed by Anatel. In practice, SNOA works as a wholesale stock exchange. The orders of each product are placed online and these orders are managed until the products are


113

finally delivered. Therefore, it is possible to manage not only the commercial aspects involved in a relationship between providers, but also to deal with operational issues such as delivery times, access denials and other difficulties that could mean excess market power by a dominant provider. SNOA also has an infrastructure georeferencing system. This allows accurate monitoring of existing supply conditions, also serving as a monitoring panel of the telecom‑ munications infrastructure in Brazil. In short, the system brings a significant number of benefits for small pro‑ viders, encouraging competition: • Transparency: all supply conditions are clear and previously available, including contractual drafts. • Equality and non-discrimination: the existing trade conditions apply to all providers buying products on the platform, so that discriminatory practices that undermine competition are avoided. • Transaction cost reduction: before the existence of SNOA, access to this type of input was expensive, discriminatory and often denied. This increased transaction costs, both because of difficult access to the sales areas of the providers and because of after-sales procedures. With SNOA, these transaction costs are greatly reduced, because the nego‑ tiation, signatures, document management, order and delivery control are made through the system.

Development and governance As already noted, SNOA is maintained and managed by the service pro‑ viders, under the coordination of Anatel. A management structure was cre‑ ated by PGMC itself. It established the Group for Implementation of the Wholesale Supply Supervisor Entity (Grupo de Implementação da Entidade Supervisora de Ofertas de Atacado e das Bases de Dados de Atacado – Giesb) in 2013 as a group created and coordinated by Anatel. The Giesb has the par‑ ticipation of all six providers having SMP as well as non-SMP companies and business associations. It is a discussion forum, where SNOA features and characteristics were defined, as well as the agency to be hired to develop and operate it (Figure 4.5). The agency hired to act as ESOA (see Figure 4.4) was the Brazilian Association of Telecommunications Resources (Associação Brasileira de Recursos em Telecomunicações – ABR Telecom).

114


FIGURE 4.5. Structure of the Group for Implementation of the Wholesale Supply Supervisor Entity (Giesb) GIESB

WHOLESALE COUNCIL OPERATORS WITHOUT SMP

OPERATORS WITH SMP

ESOA

Source: Anatel

Giesb has an executive group formed by representatives of the providers and coordinated by Anatel’s competition overseer. The group has two per‑ manent technical subgroups: GT-1, which deals with legal issues related to the entity and the apportionment of costs for the development and operation of the platform; and GT-2, which deals with the definition of requirements, architecture, operation and other technical issues. For the governance of SNOA, the ESOA10 is under the direction of a Wholesale Council formed by the providers with and without SMP. The structure of this council is defined in a balanced manner, with parity of votes between the group of companies with SMP and the group of companies without SMP. This governance model enables the contribution of both sides in a balanced manner, reducing conflicts and providing the agents with a neutral trading platform.

Stages of implementation of SNOA and general operation data After initial discussions, the project was divided into stages, so as to meet the provisions of the PGMC and to form a robust system with a consistent architecture for all participants. 10 The entity-related information is available on: . Accessed: January 14th, 2016.


115

The phases determined for system implementation were: • Phase I: system operated without online integration with providers. Existence of a single purchase interface, the web interface. In this phase, all data load determined by the PGMC is already available: ref‑ erence offers, legacy contracts, legacy circuits, and registers in general (wire centers, towers, addresses, and scope, PTT). • Phase II: automation of local BDA (wholesale databases) and online integration of these with centralized BDA; entry into production of the communication interface via web service. • Phase III: integration of legacy systems of companies with local BDA in order to complete the process of integration of the internal proce‑ dures of each SMP provider with the SNOA. Phase I of the project, which actually corresponds to the fulfillment of the objective determinations of PGMC, occurred within the period prescribed by the Regulation, on September 12th, 2013. Phases II and III, after some post‑ ponements discussed by Giesb, were fulfilled over 2015. After two years of system operation, some data can be presented. Figure 4.6 shows the growth in the number of companies able to operate the system.

apr/2015

may/2015

feb/2015

mar/2015

jan/2015

dec/2014

oct/2014

nov/2014

sep/2014

jul/2014

aug/2014

jun/2014

may/2014

apr/2014

mar/2014

jan/2014

feb/2014

dec/2013

nov/2013

oct/2013

sep/2013

FIGURE 4.6. Growth in the number of companies able to operate in the SNOA

Source: Anatel

116


Figure 4.7 shows a map with switching centers, towers and points of traf‑ fic exchange of providers with SMP. The system has about 30,000 wire cen‑ ters, 41,000 towers and 92 points of traffic exchange.

FIGURE 4.7. Wire centers, towers and points of traffic exchange available in SNOA

Switching centers

Towers

Points of trafic exchange Source: Anatel

Through December 2015 there had already been about 40,000 orders, with an average of about 1,200 orders/month. Figure 4.8 shows the evolution of the most traded product on the plat‑ form, the EILD product.11

11 There is an interesting comment to be made on this matter. Prior to SNOA deployment there had always been much discussion in the industry about the characterization of the EILD product. This product is the delivery of a dedicated circuit between two specific geographical points. When a company requests this circuit and the supplier does not have any pre-existing infrastructure, this circuit is called Special EILD, whose network deployment costs are passed on to the requester. Therefore, this is a more expensive product than the standard EILD, based on the existing infrastructure. Before the implementation of SNOA, many questioned the fact that most of the circuits were characterized as special. After implementing the system, which classifies circuits from the rules set out in the regulations, especially after April 2014, there was a significant reduction in circuits classified as special. This was seen by the industry as one of the main benefits of SNOA.


117

apr/2015

may/2015

feb/2015

mar/2015

jan/2015

dec/2014

oct/2014

nov/2014

sep/2014

aug/2014

jul/2014

jun/2014

may/2014

apr/2014

mar/2014

jan/2014

feb/2014

dec/2013

nov/2013

oct/2013

sep/2013

FIGURE 4.8. Evolution of the number of requests for EILD

Source: Anatel

SNOA also provides performance indicators for each provider with SMP in the supply of various products, indicating delivery times and order status, al‑ lowing more accurate management and transparent negotiations of each item.

CRITICISM OF THE PGMC AND NEXT STEPS After more than two years of the PGMC, the market regards it as a positive pro-competitive tool. Transparency and equality measures were positive, as they ensure non-discriminatory treatment. Moreover, they reduced the negotiation cost for purchase and delivery of products, especially in interconnection markets and EILD. Moreover, the sharing of tower passive infrastructure shows steady growth, reducing the needs for operators to invest in their own towers and thus optimizing business costs. This trend can be seen worldwide, not only for towers but also for other types of passive infrastructure, such as ducts and dark fibers, since it usually results in reduced costs for companies and opti‑ mizes the allocation of investments. On the other hand, some criticisms are still leveled at the PGMC. Despite the progress made, there has been no negotiation of the so-called access network products (full unbundling and bitstream). Unlike Europe,

118


where these products are widely traded, there has been no request or order for these products in Brazil since SNOA came into force. There may be two reasons for that. The first is that the PGMC was launched late, so entrant companies found better alternatives for construc‑ tion of access networks than traditional copper networks. This measure would have been most welcome in the middle of the last decade, but now it offers little competitive advantage. It is only attractive in cases of network extensions. The second explanation is the limited attractiveness of this solution, since the available trading conditions would not allow its further use. They require a very high minimum volume of orders, which discour‑ ages such deals. Another criticism relates to market products for transportation networks. Critics claim that the products available are not adequate to the needs of the contracting plaintiff. The EILD product, highly demanded for telephony in the corporate segment, is not suitable for broadband providers. It is alleged that the need would be for high-speed circuits (above 100 Mbps), while this product has speeds of up to 34 Mbps. Class V Interconnection and Internlink products, on the other hand, are designed to provide connectivity among telecommunications service providers to provide access to the Internet, with trading conditions far less attractive than those already found in the market. In other words, they say there are unregulated products that are cheaper for new entrants than those found in SNOA. In view of these criticisms, Anatel has argued that the regulation laying down the criteria for approval of bids is still being developed. This regulation is essential and should be based on principles and criteria used for analyz‑ ing business conditions to ensure the replicability of traded wholesale prod‑ ucts. This regulation should go to public consultation in 2016. Moreover, the agency is preparing a review in the definitions of “relevant markets” and “products”, scheduled to take place by the end of 2016, which should bring a more appropriate definition of the products and markets. Furthermore, after two years of the PGMC, other challenges have been made regarding the promotion of competition in the industry. One of the main problems, especially for Internet providers building broadband networks, is the difficult access to poles for deploying fiber optic cables. This infrastructure, mainly owned by electricity distributors, is con‑ sidered essential for building networks. In general, the prices charged by


119

these distributors are considered prohibitive and discriminatory. The mar‑ ket says they affect competition in a bad way. Likewise, the lack of transpar‑ ency in the use of poles supposedly favors groups with SMP. In addition to access to favorable business conditions, they can also monopolize the use of the poles. This requires regulatory intervention. Anatel and the National Electric Energy Agency (Agência Nacional de Energia Elétrica – Aneel) promulgated Joint Resolution 4 in December 16th, 2014, approving the reference price for the sharing of poles between electric‑ ity distributors and providers of telecom services.12 This regulation, which is still being implemented in February 2016, includes rules for the occupation of the attachment points of the poles, establishing a reference price and de‑ termining transparency obligations aimed at assuring better conditions for sharing these elements. Another issue raised as a major challenge of the competition is called competition with Over The Top applications (OTT).13 These applications usu‑ ally offer end users the same facilities as the traditional telecommunications services, albeit with different regulatory conditions. This asymmetry has allowed the growth of application providers on a global scale, while telecom providers have gradually lost the value of their business. Figure 4.9 shows the evolution of stock prices of the leading OTT applications compared with major telecommunications providers.

12 Available on: . Accessed: January 14th, 2016. 13 Over the Top (OTT) generally refers to applications for the provision of services such as audio, video calls, and instant messaging. These applications use the IP connection provided by broadband providers and offer services usually already offered by these or other traditional telecommunications providers. Examples of OTT applications are WhatsApp, Facebook, Netflix and Skype.


120

FIGURE 4.9. Evolution of the average change in stock price: OTT companies and major telecommunications companies 400%

350%

Variation compared to the value of shares in jan/2012

300%

250%

200% OTT Telco 150%

OTTs considered: Ama zon (AMZN), Facebook (FB), Google (GOOG), Yahoo (YHOO) a nd Netflix (NFLX)

100%

TELCOs cons idered: Ameri ca Movil (AMX), Telecom i talia (TI), Telefonica (TEF), Telecom Argentina (TEO), Entel (Chile), (ENTEL.SN), Oi (OIBR), Verizon (VZ) and AT&T (T)

50%

apr/16 abr/16

dez/15 dec/15

dez/14 dec/14

dec/13 dez/13

dez/12 dec/12

jan/12 jan/12

0%

-50%

Source: Public data consolidated by Anatel

This effect is called “crocodile mouth” by the financial market, because it shows that the telecommunications sector has lost economic value at the expense of companies providing OTT services. This issue is very complex and has global scope. On the one hand there is the loss of revenue from traditional services to OTT services. On the other hand, these providers generate an increase in broad‑ band traffic to their applications using the infrastructure of telecom providers. Of course customers today prefer these services over traditional services, either because of the commercial terms (often free, see Skype and WhatsApp) or because of the added features that generate value to these applications. However, we cannot yet say that there is market dominance, or any other is‑ sue justifying regulatory intervention. As a matter of fact, we are experiencing a moment of technological dis‑ ruption, where new products and new businesses emerge all the time. In this sense, regulators discuss how to ensure the so-called level playing field (equal conditions) in the relationship between traditional providers and OTT. There


121

is as yet no consensus on how to handle this issue, and therefore it is likely to pervade discussions about competition in broadband markets for a long time.

OVERVIEW OF COMPETITION IN BRAZIL TODAY After nearly 20 years of liberalization in the telecommunications market in Brazil, where there used to be a state monopoly on all services, the promo‑ tion of competition has changed the reality of the provision of telecommu‑ nications services. In particular, with regard to the provision of broadband services, we find that the monopoly conditions have been greatly reduced. The following graphs illustrate this for fixed broadband (Multimedia Communication Service – SCM) and mobile broadband (Personal Mobile Service – SMP).

FIGURE 4.10. Number of fixed broadband providers (SCM) in Brazilian municipalities (Dec/2015)

Number of SCM operators 1 2 3 4

Number of operators

Numbers of Percentage of municipalities municiplaities

5

1

1,179

6

2

1,201

22%

7

3

948

17%

8

4

783

14%

5

518

9%

6

348

6%

7

196

4%

8

100

2%

9

63

1%

10 or more

234

4%

9 10 or more

Source: Anatel.

21%

122


FIGURE 4.11. Number of mobile broadband providers (SMP) in Brazilian municipalities (Dec/2015)

Number of mobile broadband operators (Presence of ERBs 3G or 4G) Without operator 1 2 3 4 5 6

Number of operators

Numbers of municipalities

Without operator

875

Percentage of municiplaities 16%

1

1,843

33%

2

1,382

25% 12%

3

676

4

445

8%

5

345

6%

6

4

0.07%

Source: Anatel.

The data show that by the end of 2015 79% of Brazilian municipalities were served by more than one provider in fixed broadband. In mobile broad‑ band, 51% of the municipalities have more than one provider. If we pay at‑ tention to these numbers, we will notice there have been significant advances in the promotion of competition in Brazil. However, the data also show that much remains to be done The PGMC is the main regulatory tool to promote competition and it must improve its understanding of the Brazilian context. We know that the promo‑ tion of competition is not applicable to all areas, in view of the socio-economic conditions that often make it impossible to build infrastructure and provide


123

services in natural market conditions.14 Thus, a more accurate reading of the Brazilian context is needed so that we can understand in what areas we can still advance and in what areas a more specific public policy is needed.

CONCLUSION This chapter presented the PGMC, its results, characteristics, theoretical basis, and the criticism from the market, especially with regard to the promo‑ tion of competition in broadband services. As shown, the PGMC is a set of measures adopted by Anatel to pro‑ mote competition. After finding some market failures, the PGMC iden‑ tifies providers with market dominance, the so-called SMP companies, and requires those companies to abide by a set of transparency, equality, and price control obligations to reduce the barriers to the entry of new providers. The main measure implemented under the PGMC is the creation of a Wholesale Negotiation System, the SNOA. This system is a wholesale stock exchange where the necessary complements to the providers’ infra‑ structure are marketed. This platform was implemented in 2013 and it has already created a number of benefits to competition, like greater trans‑ parency, reduction of negotiation costs, and better business conditions. However, many measures still need to be implemented by Anatel, espe‑ cially in regard to prices. Finally, the challenges faced in promoting competition in Brazil are still significant. Whether in the dominance relationship between telecommunica‑ tions providers or in their relationship with dominant companies from other sectors in the supply chain, there is still much to be discussed. Therefore, this topic must continue to be addressed as in the development of Brazilian tele‑ communications, particularly the expansion of broadband.

14 The World Bank has developed the so-called Gaps Model, in which there are supply shortages and low-income regions that make it virtually impossible to promote competition. More information about the model is available at: World Bank (2011, Figure 6.2, p. 157.). See also, in this book, Chapter 13, Figure 13.4.

124


REFERENCES CAVE, M. Encouraging infrastructure competition via the ladder of investment. Telecommunications Policy, n. 30, p. 223237, 2006. MARQUES, F. O. (ed.). Código de Defesa do Consumidor, legislação dedefesa comercial e da concorrência, legislação das agências reguladoras, Constituição Federal. São Paulo: Editora Revista dos Tribunais, 2004. MOTTA, M. Competition Policy: Theory and Practice. Cambridge: Cambridge University Press, 2004. WORLD BANK. Telecommunications Regulation Handbook. 10th ed. Washington, DC: World Bank, 2011. Available on: . Accessed: May 9th, 2016.

|5|

FROM SCARCITY TO ABUNDANCE: THE DEBATE ON THE EFFICIENT USE OF THE ELECTROMAGNETIC SPECTRUM Nathalia Foditsch [ American University ]

Luca Belli [ Rio de Janeiro Law School from Fundação Getulio Vargas]

F

INTRODUCTION

or non-engineers or non-physicists, the electromagnetic spectrum is something almost metaphysical, a part of our reality that is in‑ accessible to the eyes. It is behind much of the technologies that connect us today, but how can a resource with which we are not really familiar influence our lives so much? If today smoke signals or Morse code are not used anymore, it is because the Internet and, more recently, smartphone devices connect us all using this resource. The electromagnet‑ ic spectrum is the set of all the frequencies contained in electromagnetic radiation, including X-rays, visible light, gamma rays and radio spectrum. Electromagnetic radiation involves light waves and radio waves, transmit‑ ted in the form of electrical and electromagnetic fields. They are called waves due to their oscillating nature, like a wave, whose oscillation rate – expressed in cycles per second – is what is called spectrum frequency (MUSEY; GOLZ, 2013). There is a great opportunity to increase the provision of wireless broad‑ band due to the emergence of new technologies that enable more efficient use of the spectrum. Similarly, spectrum management models can be improved, contributing to a better use of this important resource. Moreover, there are opportunities and challenges related to the development of an ecosystem that allows a wider range of stakeholders to have access to and explore wireless broadband, as we will explain next.

126

nathalia foditsch • luca belli

In the first section, we briefly describe some aspects of the development of wireless broadband. In the second, we will cover the “traditional” mod‑ els of spectrum management and then explore some aspects related to the evolution of the electromagnetic spectrum management. Next, in the third section, we analyze the “license-based” model – particularly with regards to technological neutrality and to secondary markets of spectrum – to demon‑ strate the importance of a management paradigm that also enables the use of the unlicensed model, the “spectrum commons”, which we discuss in the last section of the chapter.

WIRELESS BROADBAND AND THE ELECTROMAGNETIC SPECTRUM Since 2014 the number of online devices is larger than the world’s popula‑ tion. Moreover, in 2015 alone, mobile data traffic increased by 74%. By 2020, global mobile data traffic is expected to reach 30.6 exabytes of information every month (CISCO, 2016). Most of us are not even familiar with the scale of exabytes. Be that as it may, even if we cannot estimate with precision what the consequences of this amount of bytes will be, we can recognize that the growth of data traffic in wireless Internet is a clear trend. Concepts such as the “Internet of things” (IoT) have already been expand‑ ed to others, such as the “Internet of all things,” the “Internet of people”, the “Internet of money” and even the “Internet of food”. While everyone seems to want to coin new terms, technology evolves and the demand for data grows, particularly within the wireless Internet. Consequently, there must be a wire‑ less Internet ecosystem prepared to meet this demand. The private sector and various academics see the risk of an exponential increase in demand for data (CISCO, 2016; Latouche et al, 2013.); a kind of “apocalypse” after which telecommunications networks would no longer be able to support the demand for wireless broadband because of spectrum scar‑ city (CHAPIN, LEHR, 2011). Therefore, data explosion would accompany a real electromagnetic spectrum crisis. Others, however, argue that we need to use our networks more efficiently, promote investments in infrastructure and make use of the already allocated spectrum (CALABRESE, 2009; STAPLE; WERBACH, 2004; GARCIA ZABALLOS; FODITSCH, 2015). Moreover, some studies show that the underutilization of the allocated spectrum is a practice adopted by some and that the argument of the spectrum crisis is thus fallacious (MEHTA; MUSEY, 2015).

from scarcity to abundance

127

It is noteworthy that, according to International Telecommunication Union (ITU) data, 69% of the total population is covered by 3G networks, and 89% of the urban population and 29% of the population in rural areas is cov‑ ered.1 In Brazil, there are nearly 1.3 phones for every person in the country,2 and about 50% of the people have access to mobile Internet.3 Mobile coverage is above the world’s average, and only 5% of the population live in areas not covered by 3G networks.4 When it comes to wireless networks, spectrum can be used for other types of wireless broadband technologies, such as Wi-Fi. Pioneering research has demonstrated the possibility of using visible light to transmit data at high speed, exploring LED technology under Light Fidelity or “Li-Fi” (TSONEV; VIDEV; HASS, 2013; SCHMID et al, 2013). With regards to the total number of Wi-Fi hotspots, data shows an exponential growth in Brazil, with a 298% increase between 2013 and 2016. In 2016 there were more than 225,000 busi‑ nesses using the technology,5 which is also used in households and commu‑ nity networks.

1

Brazil ICT Country Profile. ITU Eye, 2013. Research available on: . Accessed: March 8th, 2016.

2

Available on: . Accessed: May 7th, 2016.

3

According to data from the National Sample Survey of Households – PNAD 2013, published in 2015 by the Brazilian Institute of Geography and Statistics (IBGE), 43.5% of households in Brazil have access to mobile broadband. The data has certainly changed since 2013, in other words, there is an even greater number of people using mobile broadband today. Research available on: . Accessed: February 7th, 2016. Moreover, data from the International Telecommunication Union (ITU) for 2013 shows that 51.5% of people have access to mobile broadband. Brazil ICT Country Profile. ITU Eye, 2013. Research available on: . Accessed: February 7th, 2016.

4

Available on: . Accessed: May 7th, 2016.

5

Data for the period between 2013 and 2016. Available on: . Accessed: March 7th, 2016.

128


Distinction between mobile broadband and wireless broadband Broadband wireless and mobile broadband are related concepts, but they are not synonyms. We can say that every mobile broadband is wireless, but the opposite is not true. Wireless broadband can be “nomadic” or mobile. It is nomadic when it comes from a “semimobile” connection, i.e. mobile within a limited area (e.g. Wi-Fi) that depends on the access to a fixed network. According to recent data, in 2015, 51% of total wireless data traffic was through nomadic connections (CISCO, 2016). Mobile broadband, however, is a type of wireless connection that allows mobility within a larger coverage area (e.g. 3G and 4G technologies). The term “wireless” can therefore designate both standards, nomadic and mobile.

Although the total number of devices connected to the worldwide web is im‑ pressive and despite the encouraging data from the ITU on coverage, we should note that about half of the world population has access to the Internet. Similarly, Brazil still faces challenges related to the universal access to broadband. A study published in 2016 points out that the prices of mobile broadband in Brazil are competitive in relation to other LAC countries if the purchasing power pari‑ ty is considered (CISCO, 2016). Despite this relative advantage over some of its neighbors, the price of broadband in Brazil is still far higher than it is in countries from the Organization for Economic Cooperation and Development (OECD) (GALPERIN; RUZZIER, 2011; A4AI, 2015). As we can see, there is a great opportunity to increase the provision of wireless broadband. This opportunity exists because of new technologies that enable more efficient use of spectrum as well as possible regulatory innova‑ tions. Thus, although the numbers show an increasing demand for wireless broadband in recent years, it is also necessary to understand that the current management models as well as public policies and regulations can still be improved.

Traditional models of spectrum management Spectrum management models have traditionally been classified as: (i) command and control; (ii) licensed; and (iii) unlicensed. In Brazil and in most countries, electromagnetic spectrum management has traditionally been


129

done through the “command and control” model. This model is based on the top-down control by the government of each country, through which the gov‑ ernment has the power to make allocations, assignments and allotments re‑ garding spectrum use. The licensed model, on the other hand, is an approach based on market rules, in which parts of the spectrum are assigned to differ‑ ent carriers, following the rules set by the regulators. Under the unlicensed model, there may be rules regarding the equipment to be used, but there is no licensing or assignment of exclusive rights in the use of frequencies. The command and control model is justified by the need to avoid inter‑ ference – i.e. wave superposition due to the simultaneous use of the same spectrum range – and also because of the growing demand for spectrum over the past decades. Through this model the government can control the use of the spectrum, the technologies to be used and the services to be offered. The government is the central authority of the command and control model. However, this model does not guarantee that the spectrum is used efficiently, or even whether the frequencies are effectively used. The underutilization of the spectrum is one of the main challenges of this model. In this sense, the prohibition of subletting or reselling underused bands of the spectrum is a shortcoming of the command and control model, since it does not exploit the spectrum that is not being used. An example of how this model is applied to the Brazilian reality is the expected development of a plan by the National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel) to allocate radio frequency bands for single purposes: military, telecommunications services to be pro‑ vided under the public regime and under the private regime, broadcasting services, emergency services and public safety, and other telecommunica‑ tions activities, pursuant to § 1 of Art. 158 of Law 9,772 / 97, the General Telecommunications Law (Lei Geral das Telecomunicações – LGT). Because of their limited financial and human resources, there may be an asymmetry of information from the government in regards to the regulated market in the command and control model. Thus, the command and control decisions may be suboptimal. That is, in a competitive market, competition among telecommunications companies may be a way to ensure a more effi‑ cient allocation of resources such as spectrum bands. This logic leads to the other spectrum management models discussed below.

130


The licensed model grants users the exclusive right to use the spectrum. These rights are generally granted by administrative means, through mech‑ anisms such as auctions. It is a justifiable model when there is scarcity of spectrum and when transaction costs (e.g. bidding processes) associated with the access rights are low. This model can also lead to the underutilization of the spectrum, just like the command and control model described above. An example of how this model is applied to the Brazilian context is the always costly bid for the authorization of radio frequency use, provided for in Article 164 of the LGT. The unlicensed model – or commons (common property) – is a model in which the use of the spectrum is open to everyone. Interference is avoided by using technologies that enable this sharing. As Sims (2015) explained, ITU ra‑ dio regulations do not allow the spectrum to be used without a license (SIMS; YOUELL; WORMERSLEY, 2016). Therefore, the unlicensed spectrum is also subject to some control by regulators, which allow certain devices (e.g. Wi-Fi devices) to be used. Such technologies are always changing and increasingly challenging the notion that there is scarcity of spectrum. This model is particularly justified in cases where the transaction costs for licensing cannot be warranted. The unli‑ censed model has the potential to encourage innovation, individual initiative and empowerment of local communities as it reduces the barriers to entry, allowing the testing of new spectrum uses and the creation of new technolo‑ gies (BAIG et al., 2015). According to Article 157 of the LGT, the spectrum is a limited resource and a public asset. This public nature makes the spectrum subject to regu‑ lation and requires that due attention be given to the subject. Technological developments in recent decades have allowed mobility and convergence of services, and have provoked a reflection on the need for expansion of existing services, while new services are created for the use of spectrum. Therefore, the spectrum management model is a fundamental issue. Thus, there is today an evolution of these models that aims to meet both current and future de‑ mands. The table below summarizes the management models we described, as well as their advantages and disadvantages.


131

TABLE 5.1. Spectrum management models Concept

Principal Characteristics

Disadvantages

• Regulation defines all the characteristics – such as use, technologies, and services • Dedicated spectrum

• Possibility of international harmonization • Greater control as regards universalization goals

• Sub-optimal allocation • Possible disincentive to innovate • Possible underutilization of spectrum

Market-based property rights

• Exclusive rights • Market mechanisms

• Possibility for creation of secondary markets • Possibility of establishing obligations for coverage and universalization • Adequate when there is little available spectrum and transmission costs associated with access rights are low

• High transactions costs can be problematic, especially for new entrants. • Possible underutilization of frequencies

“Commons”

• Shared spectrum

• Lower barriers to entry, encouraging new operators and technological innovation • Flexibility in the use of frequenciss for users that respect the established technical limits and equipment certifications. • Adequate in situations with little shortage of spectrum and high transaction costs

• Possíble saturation • Possible “tragedy of the commons” • Impossible to establish obligations for coverage and universalization

Varied characteristics

• Combinnes characteristics of the other models

• Flexibility to adapt to characteristics of the market

• Combining distinct models in an efficient and eficatious manner is a challenge

Non-licensed Mixed

Advantages

Centralized

Licence-based

Command and control

Model

Source: Authors based on Garcia Zaballos and Foditsch (2015, p. 14).

132


While the command and control model has been used in most countries over the past decades, there has been an evolution toward mixed models, which combine features of the three models indicated above. Below we will explain how new technologies are driving this shift toward licensed (license-based) and unlicensed (commons) models.

Toward the license-based model The technological development of the last decades showed the need to update our legal, political and regulatory frameworks concerning the Internet and the convergence of services of traditional telecommunications in the digi‑ tal environment. This worldwide trend extends to Brazil. With that in mind, a series of public debates and consultations were promoted to discuss the devel‑ opment of technologies and communication. Examples of this participatory trend are the public consultation organized by the National Agency of Civil Aviation in order to regulate the use of drones6 and the public consultation promoted by the Ministry of the Communications in the end of 2015 to dis‑ cuss how to provide telecom services.7 Another example of participatory democracy, recognized internationally, is the debate promoted by the Ministry of Justice to improve the regulatory framework of Law 12.965/2014, the Brazilian Civil Rights Framework for the Internet (Marco Civil da Internet). This Framework was approved after many years of debate. It established several principles, guarantees, rights and duties for the use of Internet in Brazil. Among these basic elements, it is worth high‑ lighting the role of net neutrality. This principle plays a fundamental role to guarantee the openness of the Internet, promoting innovation and protecting the basic rights of the users. According to Marco Civil da Internet, net neu‑ trality must be preserved, guaranteeing the equal treatment of data packages, “without any distinction of content, origin and destination, service, terminal or application” (art. 9). The principle of neutrality should be considered in the context of spec‑ trum management, but should be done from a different perspective. When it comes to the regulation of the electromagnetic spectrum, the term neutral‑ ity refers not only to net neutrality as discussed in the Brazilian Civil Rights 6 Available on: . Accessed: March 7th, 2016. 7

Available on: . Accessed: March 7th, 2016.


133

Framework for the Internet. Technological neutrality refers to the types of technology used for electronic communications services that can be provided by the operator in offering a particular service (e.g. the ability to use GSM or CDMA).8 Note that service neutrality – which can be considered part of the net neutrality and a purpose of technological neutrality – is not automatically applied mutatis mutandis in spectrum management. Moreover, this becomes particularly important when it comes to enabling and promoting the neutrality of services by the operator (e.g. license for oper‑ ating mobile broadband and other services at the same time). In Brazil, as well as in other Latin American countries like Chile, Mexico and Panama, licenses determine which services can be provided. The possi‑ bility of adopting a single license for all services – voice, pay TV and Internet – is a reality in some other countries in the region such as Argentina, Colombia, Costa Rica and Peru (ITU, 2012). In a technological convergence environ‑ ment, specifying the services that can be provided can prevent carriers from benefiting from potential efficiency gains generated by technological innova‑ tions. Specific service authorizations also represent a high regulatory burden, since carriers must have many different permits to provide a full range of services to their customers. Another example of an important discussion for the development of wireless broadband markets refers to the possibility of creating secondary spectrum markets. Countries like the United States, Australia, the UK and Germany have established some form of secondary markets under the author‑ ity of regulators (GARCIA ZABALLOS; FODITSCH, 2015). Although second‑ ary markets are not the only solution to a more efficient use of the spectrum, they can, for example, help making the underutilized spectrum available to other carriers. Article 163 of the LGT provides that “the use of radio frequency, with or without exclusivity, depends on the prior grant from Anatel and is subject to authorization under the regulation.” Article 168 states that the “authorization to use radio frequencies is untransferable without a corresponding transfer of the concession, permit or authorization to provide the service linked to them.” Article 169 says that the use permit may be terminated if a transfer is made irregularly. This lack of flexibility in the “traditional” models of spec‑ trum management can prove unable to promote effective use of the spectrum. 8

GSM is the acronym for Global System for Mobile Communications. CDMA is the acronym for Code Division Multiple Access, popular patterns of mobile phones in the world.

134


Several studies (OLAFSSON; NEKOVEE; 2007, BAE et al, 2008) highlight the benefits of flexibility in traditional forms of spectrum allocation, in order to maximize public interest. In this sense, the development of a “secondary market” for spectrum could be beneficial, promoting resale, rental or leasing of radio between private entities, in accordance with the rules established by the government. Potential advantages and disadvantages of a secondary spectrum market have already been discussed within the federal government for many years. In 2006, a study for the former Department of Economic Affairs (SEAE) of the Ministry of Finance argued that, despite all the advantages offered by the introduction of secondary markets, we must also consider the disadvantages of such model, likely to hamper and discourage its adoption, such as (i) high transaction costs, (ii) substantial increase in risk of interference between trans‑ missions, (iii) the possibility of anti-competitive practices in this market (iv) potential problems of international standardization, (v) windfall gains and (vi) the possibility that the public interest be put aside. (LIMA, RAMOS, 2006, p. 15-16).

Nevertheless, the study concluded that the better use of the radio spec‑ trum “could be achieved through the relaxation of some rules, by issuing less strict licenses, which give their holders greater autonomy for the occupation of the licensed bandwidth”. (LIMA, RAMOS, 2006, p. 37-38). As we can see, regulatory changes to promote the license-based model – such as the creation of secondary markets – can bring benefits to the develop‑ ment of wireless broadband in Brazil. There are other countries in the world that have already taken steps in this direction to promote a more efficient use of the spectrum. However, this model should be designed in such a way that anticompetitive practices are avoided.


135

Frequency usufruct titles In Latin America and the Caribbean, Guatemala is an interesting model in the creation of market mechanisms for spectrum management. The country underwent a major reform in 1996 and granted private entities the control of wireless networks. There are three types of grants: (i) one reserved for government use, (ii) one reserved for amateurs, and (iii) the regulated bands. For regulated bands, the reform created the “title of usufruct of frequency”, a market-oriented structure in which the bonds can be fully or partially leased or sold for a period of 15 years, with extensions for another 15 years. Requests are submitted to the Superintendency of Telecommunications, an entity established as a result of the Guatemalan General Telecommunications Law, responsible for the management and supervision of the use of the spectrum, registration administration, resolution of disputes related to access and use of spectrum, as well as other spectrum-related issues (GARCIA ZABALLOS; FODITSCH, 2015, p. 65).

Toward the commons model (unlicensed) As made explicit earlier, there are new ways to increase efficiency and sharing of spectrum, whether in light of new technologies or changes in the management of this resource. For this goal to be achieved, it is necessary to examine regulatory strategies that have not yet been explored in their entirety, in order to allow the use of new ways to make the spectrum more efficient. The sharing of unlicensed spectrum is already a reality in our daily lives. Examples are the use of Wi-Fi, Bluetooth, microphones and cordless phones, and other pieces of “restricted radiation equipment”.9 In Brazil, the definition of which devices fall into this category is made by Anatel. Consumers have been using Wi-Fi to “add capacity and support services in peaks of sharp and growing demand” (BENKLER, 2012, p. 5). In recent years, new technologies that are revolutionizing the more effi‑ cient use of spectrum have been developed. Among these technologies, the “dynamic spectrum access” (DSA) is essential for spectrum sharing. This is 9 Art. 163, § 2, inc. I LGT. Furthermore, Res. 506, July 1st, 2008, regulates the Radio Communication Equipment of Restricted Radiation.

136


possible because of the development of other related technologies such as software-defined radios (SDR) and cognitive radio technologies, which enable “opportunistic spectrum access”, i.e. they detect frequencies that are not being used and adjust to the frequency bands that are available. The main concern for DSA, however, is the possibility of interference, which could adversely affect the quality of service (HASSAIN et al., 2009). There are also questions as to the reliability of the service due to competition in the use of the same frequency (ALTAMINI et al., 2013), the so-called “tragedy of the commons”.10 An important example of DSA technology is the ability to exploit the TV White Spaces (TVWS), unused parts of the spectrum that exist naturally be‑ tween two ranges that are already explored and reserved for analog television. These parts are not used purposefully, in order to avoid interference, but thanks to new technologies they can be used to send wireless broadband sig‑ nals. One of the most important features of the technologies described herein is that TVWS use spectrum spaces that are already free. Thus it is not neces‑ sary to release additional frequencies to enable DSA technologies exploiting that part of the spectrum to be used. Moreover, it is important to note that such technologies can penetrate walls and provide faster connection speeds than Wi-Fi.11 That’s why the “Super Wi-Fi”12 term is sometimes used to refer to such technologies. Although the use of TVWS has already been touted as a possibility to diversify wireless broadband in Brazil since 2010,13 Anatel has not regulated its use yet. Therefore, the use of TVWS is not yet possible in the country, although using this part of the spectrum is already a reality in other coun‑ tries. The governments of Canada, United States, United Kingdom and oth‑ er European Union members are already promoting the regulation of the use of TVWS. 10 Term coined by Garrett Hardin in 1968 to explain the dynamics that occur when a common good is shared. In particular, the author analyzes the example of pastors who share a common pasture, noting that they are in a system that encourages each one to sustain their own herd without limits, leading to overgrazing and the consequent degradation of the soil. 11 Available on: . Accessed: March 7th, 2016. 12 The Wi-Fi Alliance owns the Wi-Fi trademark and products carrying the official Wi-Fi stamp are certified as being usable with other Wi-Fi-certified products. Therefore, the term “Super Wi-Fi should not be used”. 13 Available on: . Accessed: March 7th, 2016.


137

Another particularly important element, which suggests the possibili‑ ty of a true model of spectrum commons is the phenomenon of Wireless Community Networks (WCN). The development of WCN indicates the ex‑ istence of an alternative – and concretely feasible – way to address spec‑ trum management, the construction of infrastructure and the provision of broadband. In fact, wireless networks seem to be a sustainable way to deliver access to the “last mile” by providing connectivity through radio and Wi-Fi technologies. In particular, these networks exploit the unlicensed spectrum in order to create an affordable and flexible infrastructure for the last mile (FLICKENGER, 2003; DE FILIPPI; TRÉGUER, 2015). In such a model, users respect certain rules and predetermined technical standards. Thus, this model allows an unlimited number of unlicensed users to share frequency by setting the usage rights of each participant through technical standards and codes of conduct.14 In this sense, the unlicensed spec‑ trum implies true paradigm shift, stimulating the evolution of centralized communication services, developed and managed by carriers, to decentral‑ ized networks, developed and managed by new entrants or by the users. Thus, opening the spectrum is a challenge not only to the traditional model of spec‑ trum management, based on licensing, but also to the classic paradigm of centralized management of networks, operated by large companies.15 In early 2016 there was also some debate on the Long-Term Evolution technologies (LTE), already widely used in licensed frequencies. Now they can be used in unlicensed spectrum because of new technological advances. LTE in unlicensed spectrum (LTE-U)16 is intended to certain unlicensed frequency bands and can operate in conjunction with other commercial licensed wire‑ less technologies using a technique called License Assisted Access (LAA), in which a channel in the licensed spectrum of a carrier is used as the primary channel for devices that operate on an unlicensed basis. MulteFire,17 devel‑ oped in 2015, also uses unlicensed spectrum, but it does not depend on a

14 Available on: . Accessed: March 7th, 2016. 15 Ibid. 16 Based on Rel. 10/11/12 3GPP and defined by the LTEU forum. 17 See Qualcomm presentation. Available on: . Accessed: March 7th, 2016.

138


licensed primary channel. In other words, even companies that are not hold‑ ers of spectrum licenses could use this technology. In order to explore the costs and potential benefits of LTE-U and LAA technologies, the Federal Communications Commission (FCC) of the United States released in 2015 a public notice asking for more information about these new technologies.18 The debate generated disagreements, because of the opposition of divergent interests involved. Some companies argue that such technologies allow the possibility of a just coexistence with Wi-Fi, offering better efficiency by using the resource.19 Others remain skeptical of this pos‑ sibility, since such technologies do not use the listen before talk model, which minimizes the chance of interference.20 Civil society organizations argue that one should not use the principle of “technological neutrality” simplistically and they do not want a future situa‑ tion in which having access to licensed spectrum is in fact a prerequisite to using the unlicensed spectrum.21 They fear that such new technologies could jeopardize the true nature of commons of unlicensed spectrum, which can be captured by already dominant companies.

CONCLUSIONS We can not determine whether or not there will be an apocalypse of the electromagnetic spectrum, after which telecommunications networks will no longer be able to support the demand for wireless broadband. While the de‑ mand for spectrum has grown, technologies evolve and more efficient uses for spectrum appear. Regardless of the likelihood of a shortage, it seems clear to us that demand has increased. It is thus essential for Brazil to promote a solid wireless broadband ecosystem, based on efficiency and diversification of spectrum management models. 18 Available on: . Accessed: March 7th, 2016. 19 Qualcomm. Available on: . Accessed: March 7th, 2016. 20 Available on: ; and . Accessed: March 7th, 2016. 21 Available on: . Accessed: March 7th, 2016.


139

This ecosystem will only develop if proper attention to spectrum manage‑ ment is given, promoting public policies and regulatory advances in synergy with technological developments. Therefore, we propose the adoption of a “hybrid” management strategy of the spectrum, which combines elements of the traditional command and control model with new possibilities. It is possi‑ ble to use mechanisms from spectrum markets (e.g. the creation of secondary markets), as well as commons models (e.g. the use of technologies that enable dynamic access to the spectrum). This hybrid approach is an essential step in the development of wireless broadband in Brazil. If today there are new technologies that use the spectrum more efficiently, that fact should be seen as an opportunity, insofar as it enables the provision of wireless broadband to more people at more affordable prices. Nevertheless, it is important to consider the existence of a range of regulatory challenges arising from this opportunity. Technologies that use unlicensed spectrum, in particular, have a democratic potential that cannot be ignored. In order to promote the use of promising technologies such as DSA (e.g. TVWS), Li-Fi, LTEU, LAA, among others, the legal and regulatory framework should be adapted considering the best international practices. The possibility of using technologies that exploit unlicensed and licensed spectrum simultaneously, as is the case LTEU and LAA, brings some par‑ ticularly relevant challenges. Thus, as described above, access to unlicensed spectrum should not be limited to companies that already have large portions of the licensed spectrum. Rather, it is important that this precious resource is available to commercial and non-commercial new entrants. We stress the importance of promoting the principles of technological neutrality and service neutrality in order to promote greater flexibility and encourage innovation in spectrum use. This principle seems to play an in‑ strumental role in ensuring that the benefits of technological convergence can be enjoyed, by maximizing the public interest. Nevertheless, it is important that such guidance is not used simplistically, because according to what we explained, such an argument can be used in order to favor particular interests. We stress, therefore, the importance of establishing clear parameters in terms of technological and service neutrality. Throughout this text we addressed only some of the instrumental aspects for understanding the technological and market changes that are taking place with the aim of promoting legal and regulatory improvements. Similarly, it is important that any changes in relation to spectrum management models

140


are widely discussed and clearly defined, so that goods and services can be developed and offered in a legal security environment. We are sure that these are some of the steps toward reaching the universal access to broadband, the “Internet of all things” and many other innovations to come.

REFERENCES ALLIANCE FOR AFFORDABLE INTERNET (A4AI). The 20152016 Affordability Report. 2016. Available on: . Accessed: March 7th, 2015. BAE, J. et al. Spectrum Markets for Wireless Services. In: 3rd IEEE Symposium on New Frontiers in Dynamic Spectrum Access Networks, 2008 Available on: . Accessed: March 7th, 2015. BAIG R. et al. Guifi.net, the crowdsourced network infrastructure held in common. Catalunya, Barcelona, Spain Computer Networks, 2015. Available on: . Accessed: March 7th, 2015. BENKLER, Y. Open Wireless vs. Licensed Spectrum: Evidence from Market Adoption. in: Harvard Journal of Law and Technology, vol. 26, no. 1, 2012. Available on: . Accessed: March 8th, 2015. CALABRESE, M. The End of Spectrum ‘Scarcity’: Building on the TV Bands Database to Access Unused Public Airwaves. Washington, DC: New America Foundation, [sd]. Available on: . Accessed: April 27th, 2016. CHAPIN, J.; LEHR, W. Mobile Broadband Growth, Spectrum Scarcity, and Sustainable Competition. SS RN Scholarly Paper. Rochester, NY: Social Science Research Network, September 25, 2011. Available on: . Accessed: March 7th, 2015. CISCO. Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2015-2020. 2016 White Paper. Available in: . Accessed: February 7th, 2015. DE FILIPPI, P.; TRÉGUER, F. Expanding the Internet Commons: The Subversive Potential of Wireless Community Networks. In: Journal of peer production, no. 6, 2015. Available in: . Accessed: April 27th, 2016.


141

FLICKENGER, R. Building Wireless Community Networks. 2. ed. Sebastopol, CA: O’Reilly Media, 2001. GALPERIN, H.; Ruzzier, C. Broadband Tariffs in Latin America: Benchmarking and Analysis. SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, January 1st. 2011. Available on: . Accessed: March 7th, 2016. GARCIA ZABALLOS, A.; FODITSCH, N. Spectrum Management: The Key Lever for Achieving Universality. Washington, DC: InterAmerican Development Bank, July 15. 2015. Available on: . Accessed: March 7th, 2015. HASSAIN, E. et al. Dynamic Spectrum Access and Management in Cognitive Radio Networks. Cambridge: Cambridge Press, 2009. INTERNATIONAL TELECOMMUNICATIONS UNION (ITU). Regulatory Impact of Convergence and Broadband for the Americas, 2012. Available at: . Accessed: March 8th, 2016. LIMA, M.; RAMOS, M. Sobre O Uso Eficiente Do Espectro Radioelétrico. Brasilia: Ministry of Finance, 2006. Available on: . Acessed: October 29th, 2016. LATOUCHE, M. et al. Mobile Data Explosion How Mobile Service Providers Can Monetize the Growth in Mobile Data Services Through value added, 2013. Available on: . Accessed: March 7th, 2016. MEHTA, A.; MUSEY, JA overestimating Wireless Demand: Policy and Investment Implications of Upward Bias in Mobile Data Forecasts. SS RN Scholarly Paper. Rochester, NY: Social Science Research Network, 2015. Available on: . Accessed: March 8th, 2016. MUSEY, JA; GOLZ, G. The Spectrum Handbook 2013. New York: Summit Ridge Group, 2013. Available on: . Accessed: March 7th, 2016. OLAFSSON, S, GLOVER, B. and NEKOVEE, M. Future Management of Spectrum. In BT Technology Journal, 25 (2), 2007.

142


SCHMID, S. et al. LED to LED Visible Light Communication Networks. In: Proceedings of the Fourteenth ACM International Symposium on Mobile Ad Hoc Networking and Computing, p. 110. New York: ACM, 2013. SIMS M.; Youell, T.; WORMERSLEY, R. Understanding Spectrum liberalization. Boca Raton, FL: CRCPress, 2016. Available on: . Accessed: March 8th, 2016. STAPLE, G.; Werbach, K. The End of Spectrum Scarcity [Spectrum Allocation and Utilization]. IEEE Spectrum, v. 41, no. 3, p. 4852, Mar. 2004. TSONEV, D.; VIDEV, S.; HASS, H. Light Fidelity (LIFI): Towards AllOptical Networking, 2013. Available on: . Accessed: March 8th, 2016. VIECENS, MF; CALLORDA, F. La Brecha Digital En Latin America: Precio, Calidad Y Asequibilidad La Banda Ancha En La Región. DIRSI, 2016. Available on: . Accessed: March 8th, 2015.

|6|

ACCESS TO INFORMATION AND COMMUNICATION TECHNOLOGIES IN BRAZIL: A PERSPECTIVE OF THE DEMAND FOR BROADBAND Alexandre Barbosa Alisson Bittencourt FÁbio Senne Winston Oyadomari [ Centro Regional de Estudos para o Desenvolvimento da Sociedade da Informação – Cetic.br ]

T

INTRODUCTION

he rapid spread of information and communication technologies (ICT) in all segments of society is creating many opportunities for citizen engagement in an increasingly connected world. As the adoption of ICT by citizens, organizations and governments progresses towards universal access, new challenges are posed for policy makers seeking to promote digital inclusion. In addition to promoting access, it is also essential to foster the de‑ velopment of ICT skills needed for effective participation in the digital econo‑ my. Thus, greater dissemination of access to broadband Internet has created a new debate on the socio-economic impacts for connected populations. This debate has increasingly highlighted how individuals appropriate and utilize available network resources. This chapter discusses the current situation of broadband Internet access and use, with reference to the indicators produced by the ICT Households (TIC Domicílios) study (Brazilian Internet Steering Committee, 2015), conducted since 2005 by the Center of Studies on Information and Communication Technologies (Cetic.br), which is linked to the Brazilian Network Information Center (Núcleo de Informação e Coordenação do Ponto BR – NIC.br).

144

a. barbosa • a. bittencourt • f. senne • W. Oyadomari

We start with a descriptive analysis of household access indicators and then present the indicators that illustrate the extent of Internet use. Finally, we highlight the growth in Internet use through mobile phones in Brazil be‑ tween 2008 and 2014, in order to understand the role of socioeconomic vari‑ ables in the use of mobile Internet. Next, we present some proposals for an agenda of public policies for digital inclusion in Brazil. With the emergence of mobile technologies, the debate on the individual use of the web has undergone major changes: developing countries could leapfrog the fixed broadband adoption stage, producing users who have their first experience online through connections via mobile broadband. Therefore, the rapid expansion of mobile broadband requires some debate on the devel‑ opment of skills for the use of new technologies.

BROADBAND ACCESS AND INTERNET USE: CONCEPTUAL CONSIDERATIONS The issue regarding the effective use of the Internet and digital divide is related to criticism of an approach focused on access only. Among the au‑ thors who addressed the topic from this perspective are those who identify a second-level digital divide (VAN DIJK, 2005; DIMAGGIO et al, 2004), which is expressed in motivational differences or differences in skills that generate new forms of divide, even among those who have overcome the barrier of access (VAN DIJK, 2005). In the field of sociology, there are studies investigating how social stratifi‑ cation influences the how digital media are used, while this stratification also changes depending on the technical quality of technological devices through which individuals have access to the Internet and the degree of autonomy that individuals have to use such equipment (HARGITTAI, 2008, p. 940). Among the elements that permeate this debate, is the identification of a knowledge gap: those individuals with socioeconomic advantages progress more than the others. This makes certain inequalities worse over time, especially the devel‑ opment and acquisition of ICT skills (DIMAGGIO et al., 2004). However, there are those who argue that the limitations on the use of the web go beyond the knowledge gap. Compared to traditional media (such as television), using the Internet requires a more active interaction with applica‑ tions. In this sense, usage differences are guided not only by social and cultur‑ al variables, but also by technological constraints, including the complexity,

ACCESS TO INFORMATION AND COMMUNICATION TECHNOLOGIES

145

the cost of access and the features inherent to each technology. These con‑ straints might enable or restrict certain uses (VAN DEURSEN; VAN DIJK, 2013, p. 3). Whatever the approach adopted, an important prerequisite for the de‑ bate on inequalities in the use of the Internet is the idea that some ac‑ tivities are more advantageous than others for those who perform them. The challenge is to identify to what extent some activities represent any increase in the human, financial, social and cultural capital of the indi‑ viduals who perform them (HARGITTAI, 2008). Given that certain types of use of the Internet are identified as capital-enhancing, such activities represent opportunities for social mobility. A user looking for a job online, for example, would probably be better informed about the opportunities available. At the same time, if these activities are unevenly distributed, they can reinforce existing inequalities (HARGITTAI; HSIEH, 2013). This perspective is in line with recent studies that conclude that technology can both exacerbate or contribute to the reduction of inequalities, depend‑ ing on the critical elements related to the context in which this use falls (GALPERIN; MARISCAL; BARRANTES, 2014). Another field that has fueled the discussion about the unequal forms of Internet use – and also considered the increasingly important role of mo‑ bile media – is focused on new forms of political engagement. Concerned about the impact of the digital divide for the functioning of democracies, such research has shown that the effects of the Internet for political engage‑ ment are affected by demographic and behavioral differences (SHELLEY; THRANE; SHULMAN, 2006), unequal skills (MIN, 2010), or differenc‑ es in the individual’s previous interest in politics (BORGE; CARDENAL; MALPICA, 2012). In recent years, studies on ICT uses have gained more space in the lit‑ erature, but, for the most part, they examine the effects of such use at the aggregate country level. There are still very few studies that explore microdata bases disaggregated by specific variables of interest such as, for example, geo‑ graphic region, social class, gender or education level, in order to investigate how individuals and organizations effectively use (or do not use) the Internet via broadband connections (GALPERIN; MARISCAL; BARRANTES, 2014).

146


METHODOLOGICAL ASPECTS OF THE CETIC.BR HOUSEHOLD RESEARCH Before presenting data on the spread of broadband access in Brazil, we must explain how international benchmarks for measuring ICT in households are built and the advantages and limitations of international methodologies. The development of internationally comparable indicators for measur‑ ing broadband largely reflects the debate on the digital divide. Given that the Internet in its early days was seen through the lens of telecommunications in‑ frastructure policies – concerned for decades with the telephone service uni‑ versalization commitments – it was natural that the digital divide indicators initially addressed the gap between people with and without web access. The universal service paradigm was therefore linked to the idea of home access, set in binary form (DIMAGGIO et al., 2004, p. 89). Following the measurement paradigm in the telecom sector, global broad‑ band penetration indicators are often measured from administrative data, fed by public and private operators responsible for this service. Although these databases might reveal a significant picture of the existing infrastructure, they do not enable detailed insight into the perspective of the users, which would be fundamental for the creation of public inclusion policies. In 2007, the United Nations Statistical Commission endorsed a list of key indicators to be measured and compared at international level, on access and use of ICT. This list was developed by the Partnership on Measuring ICT for Development, a cross-sector initiative launched in 2004 to promote the availability and quality of data on access and use of ICT. Indicators have been developed on infrastructure and residential and business access, as well as indicators on the use of technology by individuals.1 In 2012, the International Telecommunication Union (ITU) creat‑ ed a group of experts on ICT household indicators (Expert Group on ICT Household Indicators – EGH), which has since been responsible for the on‑ going review of these key indicators on households and individuals on access and use of ICT. This process of development of international methodological standards that harmonize concepts, definitions and guide data collection allowed the 1 More information on: . Accessed: April 19th, 2016.


147

creation of comparable global indicators based on household surveys on the subject of ICT. This provided policy makers and companies with a new source of information for decision making. Thus, it is possible to gain insight on the main challenges to be faced in relation to universal access to the Internet, especially in developing countries. This chapter was written based on the results from the ICT Households survey, which aims to map access to ICT infrastructure in Brazilian house‑ holds and how these technologies are used by individuals, with a probability sample of 19,500 households in the 2014 edition (Brazilian Internet Steering Committee, 2015). Conducted annually since 2005, the target of the survey is Brazilian households and individuals over 10 years old. Data is collected in personal interviews throughout the country, covering both urban and rural areas, through structured questionnaires. The sample of ICT Households is based on the Census of the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística – IBGE) and was designed by layering clusters in multiple stag‑ es. The sample was systematically selected with probability proportional to the size of the population over 10 years old. In 2014, 36 strata were defined with different clusters by state (UF), capital cities and countryside. For nine states, the study also considered their metropolitan areas (RM). In the North Region, five states were consolidated. These strata were used for the proba‑ bilistic selection of municipalities. For the analysis presented in this chapter, we used the data from the ICT Households between 2008 and 2014. To understand the role of mobile phone in digital inclusion, we analyzed the variables of the research for the mobile individual use and activities performed by this type of user in comparison to those who also make use of other devices to access the Internet.

INTERNET ACCESS IN BRAZIL: A HISTORY OF SOCIOECONOMIC, REGIONAL AND GEOGRAPHICAL INEQUALITIES With a population of 204 million people in 2015, spread over 8.5 million square kilometers, divided administratively into five geographical macro-re‑ gions, 26 states and a Federal District, Brazil’s regional and demographic di‑ versity poses a challenge to the creation of the universal infrastructure neces‑ sary for technological development in the country.

148


Among the main challenges to the provision of infrastructure for ICT ac‑ cess in Brazil we find aspects related to the large territory of the country and the socioeconomic disparities that produce very different penetration rates. These rates depend on the geographic region, the location of households in urban or rural areas and socioeconomic status variables (class and household monthly average income). Data from the “ICT Households 2014” survey shows, for example, that while 54% of households located in urban areas of the country have access to the Internet,2 this proportion is only 22% in rural areas (Figure 6.1). Regionally, the proportion ranges from 35% in the North to 60% in the Southeast. Among the barriers that explain a lower rate in the North we can mention limited cov‑ erage by the operators. In the North, 44% of the households without Internet connection stated that the reason for that was the lack of service coverage in their area. This percentage is higher in the North than in any other region. In addition to the challenges related to the complexity of infrastructure in‑ vestment in certain regions, income inequality is also a central in explaining the differences, particularly in regions and areas that have a relatively devel‑ oped infrastructure. This is clear when we look at the Southeastern region, which has the highest proportion of households with Internet access (60%) and, at the same time, the greatest absolute number of households without access to the World Wide Web: approximately 11.5 million households. In this region, marked by the presence of the largest urban areas of Brazil, with a wide offer of technological services, the high cost of the service is the reason mentioned by 44% of the households for not having Internet access, accord‑ ing to the 2014 edition of the ICT Households (Figure 6.1).

2

In 2014, the indicator of ICT Households on home Internet access began to consider all forms of access reported by residents, without excluding households where access was available only by cell phone.


149

FIGURE 6.1. Proportion of households with Internet access, per area (2008/2014) 100

80

60

54 40

40 27 20

20 24

31 36

44

40

6

6

2008

2009

2010

50 43

Urban Total Brazil Rural

22

27

18 4

48

8

10

2011

2012

15

0 2013

2014

Source: CGI.br (2015, p. 141).

AGE AND EDUCATIONAL LEVEL: OTHER DIMENSIONS OF INEQUALITY IN INTERNET USE IN BRAZIL The debate on the use of the Internet is focused on the individual action of engaging in online activities on the web. From the point of view of the indica‑ tors, the most internationally accepted definition of an Internet user is someone who accessed the Internet at least once in the last three months for whatever reason (INTERNATIONAL TELECOMMUNICATION UNION, 2014). The first studies that addressed the extent of use reiterate how intergroup inequalities reflect disparities observed in the general population: those with less economic and social benefits, such as the poor, people with low educa‑ tional levels and people with disabilities, tend to be digitally excluded. As al‑ ready mentioned, in recent years, the debate on digital inclusion has been geared toward the analysis of different levels of inclusion, comprising types of access, skills and motivations for engaging with ICT (HELSPER, 2011). The number of Internet users in Brazil has grown steadily over the last 10 years. At the end of 2014, the country had about 94.2 million Internet users, representing 55% of the population over 10 years old. In 2008, the percentage of users was only 34% of the population in this age group.


150

Despite this growth, regional and socioeconomic inequalities persist, as evidenced by the latest editions of the survey. Southeast (63%), South (56%) and Center-West (54%) have higher percentages than those found in the North (48%) and Northeast (43%). Of individuals living in urban areas, 59% are Internet users, compared to a rate of 29% among dwellers in rural areas. Analysis of the results according to social classes shows the growth in the proportion of Internet users from classes C and DE between 2011 and 2014, although the differences between the segments remain at levels similar to those identified in previous editions of the survey (Figure 6.2).

FIGURE 6.2. Proportion of Internet users, per social class (2008/2014) percentage on the total number of individuals 100 89

94 85

94

84

97

80

80 68 72

60

40

42

73

42

77

45

80

47

78 49

54

13 2008

A B C

38

20

0

96

DE 14

13

13

14

2009

2010

2011

2012

17

2013

21

2014


The percentage of individuals who have never used the Internet was high‑ er among residents of rural areas (64%), illiterates (96%) or those who stud‑ ied up to primary education (56%), those over 45 years old (60% of those who are 45-59 years old and 82% of those 60 years old or older). According to population estimates generated based on the results of the ICT Households 2014, about 51 million Brazilians living in urban areas have never used the Internet. In the Southeast region – the region with the highest


151

proportion of Internet users in the population over 10 years old – there were about 22.4 million people who had never used the web. Among Brazilians aged 10 or more who never used the Internet, the main reasons stated for this was the lack of skill in computer use (68%), lack of interest (63%) and lack of need (48%). The lack of computer skills was the most cited reason among residents of the North (79%), among individuals with family income up to one minimum wage and in the DE class (both 71%) and among the illiterate (73%). This may indicate an association between lack of computer skills an lower educational levels. On the other hand, the lack of interest in using the Internet was the main reason pointed out by individuals of upper classes who do not access the web (83% in the A class, 75% in the B class and 82% of the individuals with family income higher than ten minimum wages who did not use the Internet). The proportion of non-users who did not use the web because of lack of interest is also directly proportional to their age: from 32% among people aged 10 to 15 years to 71% among people aged 60 or more – this is even the most men‑ tioned reason in this age group. Moreover, it is worth mentioning that 33% of Brazilians aged 10 or more who did not use the Internet considered the service expensive. This reason is even more important for the population segment with a family income of up to one minimum wage (42%), and also between those belonging to the lower classes (40% of individuals of the DE class). The North was the region with the highest percentage of the population considering it expensive (62%), far above the other parts of the country.

GROWTH OF INTERNET ACCESS THROUGH MOBILE DEVICES: MOBILE INTERNET USERS’ PROFILE IN BRAZIL The growth of connections via mobile devices and an increased reper‑ toire of applications available for mobile platforms have been identified as opportunities for digital inclusion of significant portions of the population, especially in remote areas and between the low-income classes (BOYERA, 2008; UNDP, 2012). This transformation would enable users in developing countries to have their first online experience via mobile connections. However, there are studies that show a critical perspective regarding the role of mobile devices in tackling the digital divide, highlighting the limitations of infrastructure and content inherent to this type of access

152


(NAPOLI; OBAR, 2013). From the point of view of the activities conducted online, in turn, mobile Internet users would have fewer opportunities to engage in activities associated with economic, social and cultural benefits – such as looking for a job, education and access to information. These are more difficult to run on mobile devices with limited functionality (PEARCE; RICE, 2013). This subject has been increasingly discussed in the literature on techno‑ logical factors that interfere in the forms of use, such as unequal access to ICT equipment (device divide) (PEARCE; RICE, 2013, p 724). One of the adopted strategies is the comparison between Internet users by type of equipment. This is what Donner, Gitau and Marsden (2011) did when they analyzed mo‑ bile-only users in South Africa. In Latin America, particularly in Brazil, there are few studies on the profile of Internet users who use mobile media, particularly in relation to patterns of use and online activities. One of the studies that analyzes the subject was written by Galperin et al. (2014), pointing out limitations of mobile broadband as an alternative to fixed access in Argentina. It shows adoption standards that indicate, at least in the initial stages of dissemina‑ tion of mobile technologies, a complementary trend between services – i.e. mobile technologies are mainly used by those who already have access to fixed broadband. In Brazil, through 2013 mobile phones still played a largely complementary role in Internet access, serving as a web usage platform especially among indi‑ viduals who already had household connections (BITTENCOURT et al., 2015). With the rapid growth of access via mobile devices, especially consider‑ ing the rapid adoption of smartphones, this scenario tends to change quickly, compared to the increase in the share of the population that uses the network only through these devices. It is, therefore, important to discuss the limita‑ tions of inclusion through these devices, in view of the restrictions regarding the types of accessed content and different opportunities to engage in activi‑ ties associated with economic, social and cultural benefits. Throughout the historical series of ICT Households, we can notice a growing trend in the presence of portable devices in Brazilian households. Between 2008 and 2014, for example, in Brazilian households with access to a computer, the proportion of those with a laptop grew from 10% to 60%. But the use of the Internet via mobile phones, available to 4% of Brazilians aged 10 or more in 2008, reached 47% in 2014.


153

Mobile phones have been popular devices in Brazil for some time now. The proportion of individuals over 10 years old using this type of device ranged from 67% in 2008 to 86% in 2014. In 2014, 84% of the individuals aged 10 or more had a cell phone. However, this indicator is quite determined by social class and income variables: for example, while 97% of A class indi‑ viduals had a cell phone in 2014, this proportion was 64% among members of the D and E classes. In addition to collecting information about the place where the web is accessed, a novelty in the ICT Households 2014 survey was the in‑ clusion of a question about the devices used to access the Internet. The results reveal that the cell phone is the most commonly used device to access the Internet (76%), with more mentions than desktop comput‑ ers (54%), laptops (46%), tablets (22%), televisions (7%) and videog‑ ames (5%). Considering the definition of computer of the International Telecommunication Union (ITU), which includes desktops, laptops and tablets (INTERNATIONAL TELECOMMUNICATIONS UNION, 2014), we see that 80% of the individuals aged 10 or more who use the web do it from some type of computer. Mobile phones are the most popular devices in all segments of the popu‑ lation, except for Internet users aged 60 or more, who mentioned desktops as their primary Internet access devices (78%). To examine the particularities of using the Internet through various devic‑ es, web user profiles were compared according to the devices they use to ac‑ cess the web. With this in mind, three groups of Internet users were created: • Group 1 – users of computers only, whether desktop, laptop or tablet. • Group 2 – users of mobile phones only. • Group 3 – individuals who use the Internet both on computers (desk‑ top, laptop or tablet) and via mobile phones. Among the groups described above, the largest is made up of users via computers and mobile phones (Group 3), which represents 56% of the total – indicating that such use was more of a complement to other forms of access.3 Group 2 (those who use the web via cell phones only) represents 20% of all Internet users, whereas computer-only users (Group 1) accounts for 24%. 3

In addition to computers and cell phones, the ICT Households 2014 survey presents estimates of Internet use through television, videogames and other non-stimulated devices.

154


This distribution varies among different segments of the population, with greater relevance of Groups 1 and 2 (respectively single users of computer and single users of mobile) among the lowest classes and in rural areas, as shown in Figure 6.3.

FIGURE 6.3. Proportion of Internet users per device 100

31

31 80 56

53

58

46 60

55

59

58

54

47

51 61

48

48

58

59

68

60

85 34 12

Total

26

22

16 23

REGION

27 18

25

22

2

AGE

A Class

B Class

13 45 to 59

16 to 24

10 to 15

GENDER

30

27 14

35 to 44

23

25 to 34

25

Female

19

10

47

C Class

25

Male

North

Rural

AREA

30 22

Center-East

29 17

Urban

Total

0

23

23

41

35 24

17 22

South

20

15 18

DE Class

30

Southeast

19

Northeast

20

39

5

60 or more

25

40

SOCIAL CLASS

Internet users who use the Internet both on computers (desktop, laptop or tablet) and via mobile phones (Group 3) Internet users who use the web via mobile phones only (Group 2) Internet users who use the Internet in the computer only - desktop, laptop or tablet (Group 1)


There are also significant differences in the profile of the groups ac‑ cording to social class, as shown in Figure 6.4. Among those using the Internet both on computers and mobile phones (Group 3), more than half belonged to classes A or B, whereas this socioeconomic segment rep‑ resents less than a fifth of all users who accessed the web via cell phones only (Group 2).


155

FIGURE 6.4. Profile of Internet users groups according to devices used and social class 0

Internet users who use the Internet in the computer only - desktop, laptop or tablet (Group 1)

20

11

Internet users who use the web via mobile phones only (Group 2)

Internet users who use the Internet both in the computer (desktop, laptop or tablet) and via mobile phones (Group 3)

DE Class

C Class

40

52

17

5

60

64

42

B Class

80

100

35

3

18

47

0

7

A Class


Groups 2 and 3 are composed mostly by young people, where 77% and 67%, respectively, of Internet users in these groups were up to 34 years old (Figure 6.5). Among those using only a computer, there was a greater pres‑ ence of individuals aged 35 or more (49%). On the other hand, in terms of gender, the only group with a significantly different distribution of the gener‑ al population is Group 2, in which 58% are women.

156


FIGURE 6.5. Profile of Internet users groups according to devices used and age range 0

20

Internet users who use the Internet in the computer only - desktop, laptop or tablet (Group 1)

18

Internet users who use the web via mobile phones only (Group 2)

17

Internet users who use the Internet both through a computer (desktop, laptop or tablet) and via mobile phones (Group 3)

14

15

40

60

80

17

19

22

33

28

27

25

10 to 15 years

16 to 24 years

25 to 34 years

35 to 44 years

45 to 59 years

60 or more

14

18

100

8

8 1

11 4


The results show, therefore, that the devices used to access the Internet are associated with very different user profiles. In short, those who use the web exclusively via cell phone (Group 2) are generally younger, of low or middle social classes and predominantly female. On the other hand, Internet users that use the web only via computers (Group 1) are comparatively old‑ er, although mostly from C or DE classes. Finally, Group 3 is predominantly young and is the only group where more than half of the users are from A or B classes. Thus, further study on access to certain types of devices has an important implication on the opportunities associated with this use, in view of the in‑ equalities observed in terms of age and social class. Since this is the first edition of the ICT Households survey in which in‑ formation about devices was collected, it is not possible to identify trends in usage patterns. However, based on changes observed in the last ten years, we can say that the expansion of Internet use via mobile phones makes room for new ways to use the web.


157

ICT Households 2014 also shows differences in how the activities are carried out, depending on the type of device used to access the Internet. Considering the nine most popular activities by type of device used, we notice that the users who access the web both via computers and cell phones perform these different activities more often than other web users. On the other hand, users who use the Internet only via computers or only via cell phones, that is, via a single type of device, perform these activities less often (Figure 6.6).

FIGURE 6.6. Proportion of Internet users by activities most frequently mentioned per device 0

10

20

30

40

50

60

Sending and receiving emails

87

31

78

63

55

74 58 69 57

44 46

33

84

76

47 42

Listening to music online

74

64

56

42

Watching movies or videos on websites like YouTube or Netflix

91

76

61

41

Finding information about products and services

100

67

47

Sharing content like text, images or videos online

90

83

60

Participating in social networks like Facebook, Orkut or Google+

Downloading music

80

60

Texting via Facebook, Skype, WhatsApp etc.

Reading newspapers, magazines or news

70

46

67 53 64 51

45

62

Total Internet users Internet users who use the web in the computer only (desktop, laptop or tablet) (Group 1) Internet users who use the web via mobile phones only (Group 2) Internet users who use the web both on computers (desktop, laptop or tablet) and via mobile phones (Group 3)


158


Among users who access the Internet exclusively via computers, some activities stand out: sending and receiving e-mails, searching for information about products and services and reading newspapers, magazines or news. Those who only browse the web via mobile phones, in turn, perform the fol‑ lowing activities more often: texting, social networks and content sharing. Analyzing this information and observing differences in usage patterns among these groups shows that, despite an increase in the use of the Internet via mobile phones, this does not mean the end of the historical socioeconomic inequalities in Internet usage in Brazil. The inequalities observed in the indi‑ cator’s behavior reinforce the importance of improving the research on the de‑ vices so we can better understand the social and economic impacts of the ICT.

RELEVANT ASPECTS TO CREATE A DIGITAL INCLUSION PUBLIC POLICY AGENDA IN BRAZIL Below we highlight relevant observations, based on the results of the ICT Households surveys, to the debate on public policies in Brazil. To effectively achieve universal access, we must act on several fronts, in view of the com‑ plexity of the Brazilian socioeconomic scenario: • Targeted actions are needed to expand access and reduce regional and social inequalities. The difference in the proportion of connect‑ ed households and users and between different regions is significant. This shows the need for specific policies at the regional and local level. • The majority of rural areas remain disconnected from new technolo‑ gies and need to be assisted by targeted policies. • In urban areas the situation is quite different, with the highest pro‑ portion of households with Internet access and more users. On the other hand, although residents of most urban areas are already digi‑ tally included, it is important to consider that, in absolute numbers, there is still a long way to go, especially in view of economic and social inequalities. And it is precisely the Southeast region, with the highest proportion of connected households and Internet users, that has the largest absolute number of disconnected households. • The acquisition of ICT skills is also a topic that deserves attention in elaborating inclusion strategies. Most people who declared they were not users of ICT were also those with a lower educational level.


159

• Another element that needs to be taken into account in the formulation of public policies on access to ICT is a generational issue that also re‑ lates to the acquisition of skills and abilities. Many in the elderly popu‑ lation are not Internet users, neither via computers nor cell phones. • It is important to conceive, develop and enhance public policies to expand the provision and reduce the price of broadband, as well as subsidies for the purchase of ICT equipment in low-income families – as these are relevant factors that can directly impact digital inclusion of this significant part of the population. • Finally, we must always keep mobility in mind. Today, there is a greater penetration of tablets, laptops and mobile phones among Brazilians. Although the incorporation of new equipment at an early stage is more a complementary form of access, we must recognize the substantial changes in public policies that mobile devices require. Above all, we must recognize the opportunity that mobile media rep‑ resents in terms of access, but not disregard its limitations regarding opportunities and tangible benefits to individuals who rely solely on this type of device. For the formulation of a public policy agenda to significantly increase the number of users and connected households, it is essential to focus on the particularities of access and use. Addressing the variables taking into account their complexity is a central challenge for the coming years.

REFERENCES BITTENCOURT, A. et al. Inclusão digital e mobilidade: uma análise do perfildos usuários de Internet móvel no Brasil. Proceedings of the 9th CPRLatamConference. Cancun: [s.n.]. 2015. BORGE, R.; CARDENAL, A. S.; MALPICA, C. El impacto de internet en la participación política: revisando el papel del interés político. ARBORCiencia, Pensamiento y Cultura, v. 188, n. 756, p. 733750, jul./aug. 2012. BOYERA, S. White Paper on mobile web for social development, World Wide Web Consortium, 2008. Available on: . Accessed: December 23rd, 2015. COMITÊ GESTOR DA INTERNET NO BRASIL (CGI.br). ICT Households and Enterprises 2013 – Survey on the use of Information and Communication

160


Technologies in Brazilian Households and Enterprises. São Paulo: CGI.br, 2014. Available on: Accessed: October 29th, 2016. ______. ICT Households 2014 – Survey on the use of Informatioon and Communication Technologies in Brazilian Households. São Paulo: CGI.br, 2015. Available on: . Accessed: October 29th, 2016. DIMAGGIO, P. et al. From Unequal Access to Differentiated Use: Literature Review and Agenda for Research on Digital. In: Neckermann, K. Social Inequality. New York: Russell Sage Foundation, 2004. p. 355400. DONNER, J.; Gitau, S.; MARSDEN, G. Exploring Mobileonly Internet Use: Results of the Training Study in Urban South Africa. International Journal of Communication, vol. 5, p. 574 597 2011. GALPERIN, H.; CALLORDA, F. Banda ancha móvil: ¿Complemento o sustituto? Uso y sustitución de la banda ancha en Argentina: Un análisis apartir de microdatos. In: Proceedings of the 8th CPRLatam Conference. Bogotá: [s.n.]. 2014. GALPERIN, H.; MARISCAL, J.; BARRANTES, R. The Internet and Poverty: Opening the Black Box. Victoria: DIRSI, 2014. Available on: ;. Accessed: Dec. 23rd, 2015. HARGITTAI, E. The Digital Reproduction of Inequality. In: GRUS KY, D. Social Stratification. Boulder, CO: Westview Press, 2008. p. 936,944. HARGITTAI, E. HINNANT, A. Digital Inequality: Differences in Young. Communication Research, Vol. 35, no. 5, Oct. 2008. HARGITTAI, E.; HSIEH, YLP Digital Inequality. In: Dutton, WH Oxford Handbook of Internet Studies. [Sl]: Oxford University Press, 2013. p. 129150. HELSPER, E. Digital Disconnect: In Issues of Social Exclusion, Vulnerability and Digital (Dis) engagement. Perspectives of Web 2.0 for Citizenship Education in Europe. Brno, Czech Republic: [sn]. 2011. INTERNATIONAL TELECOMMUNICATION UNION. Manual for Measuring ICT Access and Use by Households and Individuals. [Sl]: ITU, 2014. MIN, SJ From the Digital Divide to the Democratic Divide: Internet Skills, Political Interest, and the Digital Divide in Second Level Political Use Internet Journal of Information Technology & Politics, v. 7, no. 1, p. 2235, February 2010. NAPOLI, P.; OBAR, J. leapfrogging Mobile and Digital Divide Policy: Assessing the limitations of mobile Internet access, 2013. Available on: . Accessed: January 11th, 2016. PEARCE, E.; RICE, E. Digital Divides From Access to Activities. Journal of Communication, n. 63, p. 721,744, in 2013. SHELLEY, MC; THRANE, L.; SHULMAN, S. Generational Differences in Information Technology Use and Political Involvement. International Journal of Electronic Government Research, v. 2, no. 1, p. 3653, Jan./Mar. 2006. UNDP. Mobile Technologies and empowerment: enhancing human development through participation and innovation, 2012. Available at: . Accessed: Jan 11th, 2016. VAN DEURSEN, A.; VAN DIJK, J. The digital divide shifts to differences in usage. New Media and Society, vol. 0, no. 0 p. 120, 2013. VAN DIJK, J. The deepening divide: Inequality in the information society. London: Sage, 2005.

|7|

POLICY AND REGULATION: ACHIEVEMENTS AND CHALLENGES Flávia Lefèvre [ Proteste ]

T

INTRODUCTION

he establishment of public policies to build infrastructure in a democrat‑ ic fashion, including all parts of a country, is the first step to the achieve‑ ment of universal broadband Internet access, ensuring that everyone can join the world of technology, communication, education, and culture. But infrastructure alone is not enough if the government does not adopt public policies and regulatory positions that actually give citizens access to services supported by high-speed networks for Internet access. Therefore, we will address the rights established by the Brazilian Civil Rights Framework for the Internet (Marco Civil da Internet – MCI – Law 12,965/2014), especially the right to net neutrality. We will see how the regu‑ latory process will be fundamental for this new law to become more effective in view of the objective to democratize the access to Internet services. In this chapter, we present the MCI mechanisms that grant the Internet the same typical guarantees as other public services or utilities – such as elec‑ tricity, water, gas, among others – and that should play a major role in elabo‑ rating the rules that will regulate the MCI.

CONTEXT OF REGULATORY POLICY Decree 7,175/2010 established the National Broadband Program (Programa Nacional de Banda Larga – PNBL). Despite having outlined important mechanisms for the development of infrastructure in order to ensure access to broadband Internet, it trod a different path than what was determined in the General Telecommunications Law (Lei Geral das Telecomunicações – LGT) of 1997. This is because the decree did not impose goals regarding universal access and continuity for broadband, allowing

164

flávia lefèvre

investments in high capacity networks to be made only in the regions that concentrate corporate and high-income consumers. Therefore, investments were made according to the companies’ economic interests rather than a digital inclusion policy. This dynamic aggravated broadband Internet access inequalities in Brazil. This subject is further discussed in Chapter 3 of this book. But here we can make some remarks regarding art. 65, § 1, of the LGT, since it pro‑ vides that essential services cannot be delivered only on a private basis. They should also be under the public regime. Art. 18, LGT, also determined that the Executive Branch, by decree, has the power to extend the public regime to essential services. One of the main objectives of the public regime is to ensure that the government can impose universal service goals, prioritizing regions and set‑ ting schedules. Another goal is to enable the use of public resources, such as the Universal Service Fund for Telecommunications Services (Fundo de Universalização dos Serviços de Telecomunicações – Fust), so that the country’s localities that are not attractive to the private sector – due to the need for large investments and concentration of low-income consumers – can benefit from public investments. Under the private regime, however, the government has limited regulato‑ ry powers. It cannot impose obligations and schedules for private investment nor can it establish tariffs and subsidies, which would provide easier access to the Internet. But so far the government has chosen to forgo this right and the result is an undemocratic distribution of infrastructure, as revealed in 2015 by the Traffic Measurement System (Sistema de Medição de Tráfego – Simet), a project carried out by the Brazilian Network Information Center (Núcleo de Informação e Coordenação do Ponto BR – NIC.br) – the executive arm of the Brazilian Internet Steering Committee (Comitê Gestor da Internet no Brasil – CGI.br). This distribution of investments primarily reflects the economic interests of private companies operating in the market (Figure 7.1).

policy and regulation

165

FIGURE 7.1. Speed of the Internet access: broadband, fixed, download, 2014

Up to 512 Kbps From 512 Kbps to 1Mbps From 1 Mbps to 5 Mbps From 5 Mbps to 10 Mbps From 10 Mbps to 35 Mbps From 35 Mbps to 50 Mbps 50 Mbps or more All speeds

Source: Simet.

This reality has led dozens of entities focused on consumer protection and the right to communication to join forces in the Banda Larga é um Direito Seu! Campaign.1 This initiative fights for the adoption of public policies in line with what determines the LGT, to ensure faster and more democratic distri‑ bution of fiber optic infrastructure in Brazil.2 In other words, it is important to recognize that isolated infrastructure policies do not lead to significant social and economic outcomes. Studies in‑ dicate that, to achieve the goal of universal access to the Internet, it is es‑ sential to combine policies for the expansion of infrastructure with pro-ac‑ cess campaigns (GALPERIN; MARISCAL; BARRANTES, 2014). Otherwise, robust private and public investment necessary for a greater penetration of 1 Available on: . Accessed: December 22nd, 2015. 2

See also Chapter 3 of this book.

166

flávia lefèvre

infrastructure, which imply high marketing prices, may not reach the target: the country’s poorest populations. Therefore, we must combine policies targeted on infrastructure deploy‑ ment (more related to the Ministry of Communications) with local policies to promote effective access to the Internet: programs such as Digital Cities, telecenters and other training programs for citizens. Teachers, schools and universities should also be included. These policies should be enforced in the states and municipalities by the Ministries of Education and Culture and Science, Technology and Innovation. It is also fundamental to involve the Ministries of Planning and Finance in ensuring access to services like provid‑ ing distance education, scheduling medical appointments, receiving remote medical diagnosis, paying taxes, filing incident reports, participating in public consultations and more. These services depend on broadband infrastructure so we must create subsidy mechanisms to allow low-income citizens to have access to them. The MCI applied in conjunction with the LGT and the Consumer Protection Code (Código de Defesa do Consumidor – CDC) are very powerful legislative tools for policy makers and regulators, as explained in the next section. In this context, interpretation, application and regulation of the MCI is essential to meet the Internet access objectives, in order to achieve economic, social and cultural development.

THE MCI AND THE ESSENTIAL AND UNIVERSAL NATURE OF THE INTERNET CONNECTION SERVICE The MCI gave Internet access service the status of essential public service. Art. 7 of Law 12,965/2014 established that Internet access is essential to the exercise of citizenship. It enhanced the provisions of articles 1, 2, 3 and 4, which defined the fundamentals, principles and objectives of the law, enhanc‑ ing the provisions that deal with universal Internet access for everyone. The public nature of the networks and the fulfillment of their social function of ensuring Internet access to everyone are essential principles for formulating public and regulatory policies. Admittedly, in the case of Internet access service, we cannot properly speak of a public service as defined by administrative law,3 according to which 3

“According to Celso Antonio Bandeira de Mello […] we define public service as any utility or activity to be enjoyed by the population, provided by the government or by its grantees,


167

the State is organized in cooperation between several services coordinated and controlled by governments and provided by the government or by private entities that receive the right to exploit them (MEDAUAR, 1997, p. 329). Odete Medauar (1997, p. 83), when addressing the matter, proposed that, to know “when and why an activity is considered a public service”, we must bear in mind “the plan of the ruling political concept, the plan of the State and its role”. According to her, there is a “core of public services” such as water, telecommunications and electricity; however, specific purposes imply that we consider other activities to be public services aimed at providing benefits to the population. Thus, we defend the public dimension of the Internet access service, which is covered by post-modern administrative law that emerged as a re‑ sponse to social, economic and contemporary political changes. According to this concept, the State must have new responsibilities “for the protection of the essential values and for the control and repression of abuses resulting from the concentration of economic power and the development of a consum‑ er society” (JUSTEN FILHO, 2005, p. 17). In this sense, the MCI was drafted considering the compulsory and urgent recognition of the high degree of essentiality of the Internet access service and its important role in the satisfaction of social, cultural, commercial and industrial needs, both public and private. These considerations are revealed in the world of the Internet. They did not exist when the Federal Constitution was promulgated, in 1988, that is why this service is not expressly contem‑ plated there, as is the case with other services such as electricity, gas and telecommunications. Incidentally, it is important to note that § 1 of art. 9 of the Constitution provided that essential services can be established by law, and this is what the MCI did. It is true that Internet access service is considered a value-added service, in accordance with art. 61 of the LGT, and Rule 04/1995, approved by Regulation (Portaria) 148/1995, of the Ministry of Communications and, therefore, independent of delegation or authorization to be delivered. under a public law regime instituted in favor of interests defined by the legal system. […] The provision is done by the Public Administration (bodies) or its delegates (joint stock company, public company), or even its concessionaires and licensees (private companies). […] Public services are established by the Constitution and managed by the Union, Member States, the Federal District and the municipalities, which are responsible for their regulation, enforcement and control” (MEIRELLES, 1992).

168

flávia lefèvre

However, it is also true that the MCI assigned to federal, state and munici‑ pal authorities a series of guidelines for their activities intended to develop the Internet in Brazil. These guidelines included the promotion of rational man‑ agement, expansion and use of the Internet, with the participation of the CGI. br, government, academia, civil society and companies. The law also ruled that the State must periodically formulate and promote studies and set goals, strategies, plans and schedules, for the use and development of the Internet in Brazil. In other words, it established a system of social control for the reg‑ ulation of Internet access services in a participatory and democratic manner, reserving to the State the role of normative and regulating agent, so it can play its supervisory role, in accordance with art. 174 of the Federal Constitution. Based on André Demichel’s doctrine (1978), we understand that a public service does not require the exclusive application of public law, but it is a suf‑ ficient condition to justify certain concepts of administrative law. Therefore, Internet access service, regardless of its special characteristics, should receive treatment similar that accorded to essential public services, as classified by the classical theories of administrative law.

ESSENTIAL PUBLIC SERVICES AND THEIR FEATURES As mentioned above, art. 7 of the MCI stated that the Internet is essential for the full enjoyment of civil rights. And so it should be, since today there are services that can only be accessed online, like obtaining some legal docu‑ ments, filing the income tax, drawing up incident reports, applying for social programs, among others. The same article assigned a series of rights to users, rights that coincide with typical principles of public services established in the Constitution, such as water, gas, telecommunications, among others. For example, the Concessions Law (Law 8,987/1995), which deals with the delegation of pow‑ ers to explore public services, states that users should receive appropriate ser‑ vice (art. 7). The term “appropriate service” (art. 6) is defined as the service that “fulfills the conditions of regularity, continuity, efficiency, safety, moder‑ nity, generality, and courtesy in its provision, at reasonable rates.” Thus, we can conclude that the MCI, when dealing with user rights, applied typical principles of public services to the regulatory system of Internet connection, as we can see in Table 7.1.


169

TABLE 7.1. Internet users’ rights and public service principles Internet users’ rights

Public service principles

Not terminating the connection, except for default

Regularity and continuity

Maintaining the contracted quality of the Internet connection Regularity and suitability Inviolability and secrecy of the flow of Internet communications, except by court order, under the law Not supplying personal data to third parties, including connection and access records, except with free expressed and informed consent or in the cases provided for by law

Safety

Neutrality

Equality and neutrality

Source: Author’s elaboration.

When it comes to essential public services, one of the most important con‑ sequences is the enforcement of the principle of universality and continuity, as the recognition that the activity should be accessible to all – from the rich‑ est to the poorest – equally and without discrimination, regularly and without interruption. Essential public service users’ rights are also protected by the CDC (art. 22), as stated by art. 7 of the MCI. Adequate provision of service in accordance with the CDC (art. 22) also implies another important principle – moderateness, according to which the rate should be as low as possible, guaranteeing the coverage of costs related to services provided in accordance with minimum quality standards. Since this is a service the prices of which are not regulated, we understand that user protection against abusive prices should be supported by the CDC. We should not forget that the government was assigned guidelines and tasks to ensure access to everyone. That is, even if we cannot properly speak of low rates, the fact is that public bodies should be mobilized to ensure that the prices are affordable. First, through public policies for infrastructure im‑ provement – in this case the State, through the National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel), regulates the whole‑ sale rate charged for network capacity. Second, by creating subsidy mecha‑ nisms to serve the areas where there is not enough infrastructure to meet the demands of society. Thus, there are two major challenges ahead: 1. drafting the presidential decree that will regulate the MCI so that, in the exercise of regulatory power, pursuant to art. 84, IV, of the Federal

170

flávia lefèvre

Constitution, the Presidency of the Republic take measures to assure the rights established by the MCI, providing tools for the government to ensure compliance with the guidelines imposed by law; 2. promoting the participation of civil society in the field of social control, formulating new rules that may be established and claiming their rights.

THE REGULATION OF NEUTRALITY AS A GUARANTEE OF UNIVERSAL ACCESS In this environment, net neutrality plays a key role. This concept emerged as a reaction to anti-competitive practices that discriminate and prioritize net‑ work data traffic. The concept of neutrality is intended to ensure equal treat‑ ment in the transmission, switching and routing of any data packet. Therefore, we are talking about a legal tool the definition of which is not absolute and depends on what goals society has. In Brazil, the MCI has clearly outlined the concept of neutrality, including only two cases in which it accepts discrimination or degradation of traffic, or the breach of neutrality, stating that: § 1 – The discrimination or degradation of traffic will be regulated under the exclusive competence of the President of the Republic provided for in item IV of art. 84 of the Federal Constitution, for the faithful execution of this Law, considering opinions of CGI.br and the National Telecommunications Agency, and can only result from: I – technical requirements necessary for the proper provision of ser‑ vices and applications; and II – prioritization of emergency services.

In other words, even though we are talking about a legal tool, it is certain that verification of fulfillment of the neutral network management obligations is of a technical nature and should be guided by the parameters established in the MCI, which defines the Internet as “the system consisting of the set of log‑ ical protocols, structured worldwide for public and unrestricted use, in order to enable communication of data between terminals through different net‑ works” (MCI, art. 5, para. I). Thus, in addition to infrastructure, neutrality is one of the main instruments to achieve universal Internet access, to the extent that the benefits from such access mainly result in greater scope to content of all kinds – educational, cultural, and others – so that the user can access any service or content publicly available on the Internet without restriction.


171

Neutrality is a principle that aims to preserve the open nature of network architectures and values like democracy, freedom of expression, free flow of information, privacy, competitive environment, innovation, consumer rights, and other fundamental rights. The technical aspects for the verification of compliance with the obligations of non-discrimination on the Internet will depend on the scope of the concept of neutrality that is set in the forums of the debate: regulatory agencies and the Judiciary Branch.

The practice of zero rating This being the case, it is important to examine widely practiced business practices that may be infringing the principle of neutrality, which is why we will address the topic of zero rating or sponsored access. It is the free offering of content and/or applications provided by companies that usually associate with Internet connection service operators. There are diverging views with respect to such practices adopted by many suppliers of Internet service providers in several countries. Countries like Canada, Chile, the Netherlands, Slovenia and Norway have laws prohibiting zero rating, precisely because they understand that the prac‑ tice has the potential to infringe neutrality and encourage anti-competitive behavior. In the United States, although there is no express prohibition to zero rating, the Federal Communications Commission (FCC) has promoted regulatory actions of significant importance in order to ensure the neutrality of the network, such as the inclusion in early 2015 of broadband in their area of competence, under strong protests from business groups working in the field of telecommunications. On June 30th, 2015, the European Parliament, the Council and the European Commission announced an agreement4 on the essential elements of a single telecommunications market, which introduced rules to ensure an open Internet in that market. The document contains rules to ensure the prin‑ ciple of net neutrality, preventing blocking or throttling of online content, applications and services across the European Union, contributing to a single market, by reducing its fragmentation. It was established that every European should have access to open Internet. Every content and service provider must also treat all traffic equally, 4 Available on: . Accessed: December 22nd, 2015.

172

flávia lefèvre

banning paid prioritization. Zero rating is clearly mentioned in the sense that this practice may not involve blocking competing content. The agreement seeks to promote a wider range of digital service offerings, encouraging digital inclusion. The agreement also brings clear provisions to prevent the zero rating from affecting consumers’ power of choice by imposing regulatory authorities a duty to monitor and ensure compliance with the rules.

Limited service plans and zero rating In Brazil, the practice of zero rating has been widely used by mobile oper‑ ators in plans that establish data packets with limited data volumes to be used monthly by the consumer. In such cases the connection provider does not deduct the volume of data corresponding to specific applications or content from the monthly data volume contracted for. That is, the provider limits for some data types, but not others. This practice would not seem to violate neutrality, as long as consumers were able to access other applications and content online, even after the data packet ran out, albeit at a lower speed. So one cannot speak of discrimination prohibited by the MCI. However, the practice of zero rating is associated with service plans with low data volume limits per month – 200 Mb to 600 Mb – and when the data limit has been reached the connection provider maintains access to certain applications but blocks access to the rest of the Internet. Thus, if the zero rating, according to our understanding, does not violate net neutrality while the contracted data volume has not been reached, but when it has and the provider allows access to certain applications or content and blocks all the rest, non-discrimination obligations and blocking prohibi‑ tion are violated. This what the operators think too. While drafts of the MCI were under consideration, they published a booklet that they distributed to members of Congress. In this booklet they pointed out art. 9 as one of the main problems of the bill. Operators said that the marketing of packets with limited data vol‑ ume and the practice of zero rating would hurt neutrality. This is what they claimed: Article 9 of the MCI bill determines that access providers should deal with any form of data packet in an equal manner, without any service-related distinction. As a result of this determination, the sup‑ ply of services differentiated by their speed is the only thing that


173

can be marketed. In this type of offer, all packets receive the same treatment from the network and only the speed at which they are delivered to the user may vary, depending on the contracted speed. All other types of offer, some of them currently marketed, would be prohibited by the MCI, including offerings based on the data vol‑ ume consumed by the user (LEVY, 2012).

The MCI was signed on April 23rd, 2014, and became effective in June of the same year. Art. 9 was maintained. Despite the assertions transcribed above, companies continued to sell plans with meager monthly data limits. When the limits are reached, they reduce the speed of provisioning and maintain access to the Internet. As of January 2015, claiming that this business model was no longer profitable, they changed their practice. They then adopted the zero rating to pro‑ vide access to applications and content provided by companies with which they have some kind of partnership, as is the case of services like Facebook, WhatsApp and Twitter, and blocked all other types of content or applications. That is, they have adopted a practice that is contrary to net neutrality guar‑ antees, according to what they stated in the booklet distributed to members of Congress. When we analyze this practice, the first thing we can say with absolute certainty is that the practice of zero rating does not fit in any of the cases pro‑ vided for by the MCI, considering the exceptions that allow the breaking of neutrality (§ 1, art. 9), since these cases are not due to technical issues nor do they have to do with emergency situations. This is a business model geared only towards meeting the commercial interests of the operators. However, the dispute continues regarding the practice of zero rating, and the breach of neutrality remains. Therefore, it is important to rescue the Decalogue of Principles set forth by CGI.br – Resolution 2009/2003 – which, when dealing with neutrality, stated the following: “Filtering or traffic privileges must comply with technical and ethical criteria. Political, commercial, religious, cultural, or any other form of discrimination or favoritism are not allowed”.5 There are also some great contributions to the analysis of the zero rat‑ ing practice in the consensus established in the Declaration of São Paulo,6 5

Available on: . Accessed: February 12th, 2016.

6

Available on: . Accessed: February 12th, 2016.

174

flávia lefèvre

the result of the Global Multistakeholder Meeting (NETmundial)7 held in São Paulo in April 2014, with the participation of some 110 countries (GETSCHKO, 2014, p. 13). The document addresses principles for Internet governance like: • Unified and unfragmented space: The Internet should continue to be a globally coherent, interconnected, stable, unfragmented, scalable and accessible network-of-networks, based on a common set of unique identifiers and that allows data packets/information to flow freely end-to-end regardless of the lawful content. • Open and distributed architecture: The Internet should be preserved as a fertile and innovative environment based on an open system architecture, with voluntary collaboration, collec‑ tive stewardship and participation, and upholds the end-to-end nature of the open Internet, and seeks for technical experts to resolve technical issues in the appropriate venue in a manner consistent with this open, collaborative approach. • Enabling environment for sustainable innovation and creativity: The ability to innovate and create has been at the heart of the remarkable growth of the Internet and it has brought great value to the global society. For the preservation of its dynamism, Internet governance must continue to allow permissionless innovation through an enabling Internet environment, consistent with other principles in this document. Enterprise and investment in infrastructure are essential components of an enabling environment. Thus, from the perspective of internationally accepted Internet gover‑ nance principles and in accordance with the MCI, we understand that the practice of zero rating associated with plans with restricted access to certain content and applications and blocking all the rest of the web violates neu‑ trality, causing serious losses to the Brazilian society both economically and socially, creating barriers to digital inclusion.

Economic impacts of the zero rating 7

Available on: . Accessed: July 30th, 2015.


175

To evaluate the economic impacts of the zero rating practice, it is im‑ portant to consider that telecommunications infrastructure companies with significant market power (SMP)8 and that act as Internet access providers are associated with suppliers of content and applications, enhancing anti-compet‑ itive effects and discouraging innovation by small and medium enterprises. Data released by Alexa9 in 2015 reveals the ten most visited websites in Brazil, in the following order: google.com.br; facebook.com; google.com; you‑ tube.com; uol.com.br; globo.com; live.com; yahoo.com; mercadolivre.com.br and wikipedia.org. The adoption of the zero rating practice results in a concentration of eco‑ nomic power in the hands of large transnational groups. Why? This practice combines the market power of content and application suppliers – many of which are multinational companies – with the market share of leading telecom operators and Internet connection providers in Brazil, also mostly multinational companies. This vertical integration in the provision of Internet access services and delivery of applications and content affects innovation, freedom of speech, the free flow of information, cultural diversity, the economic development of low-income classes and, as a result, it may jeopardize democracy. The justification for the practice of zero rating, alleging that these plans are geared towards low-income consumers who cannot afford unlimit‑ ed plans, is unsustainable. A study conducted by Digital Fuel Monitor in European Union countries and the OECD in 2014 showed that in countries that allow this practice, 3G and 4G access plans costs increased significantly throughout 2014, especially among operators offering zero rated video services […]. In countries that have not adopted the model a tendency of local opera‑ tors to increase the data allotments of their customers was observed, without changing the price […] (SOARES RAMOS, 2015, p. 5).

In Brazil, as reported in recent years by the International Telecommunication Union (ITU), mobile broadband service contract prices are among the highest

8

See Chapter 4 on the PGMC.

9

Available on: . Accessed: April 20th, 2016.

176

flávia lefèvre

on the planet.10 Zero rating can impact service prices of Internet access, since operators pay Facebook to be able to offer this service. That could not be any different. Markets concentrated in the hands of a few powerful companies result in their having little incentive to reduce prices and improve quality. In these markets, the concentration of economic and political power may result in higher prices than in markets with greater com‑ petition and also facilitate the adoption and use of zero rating practices. In Brazil, the three largest groups in the broadband market have 77.9% of gross revenues and 86.5% of broadband accesses (Table 7.2).

TABLE 7.2. Telecommunications market share (third quarter 2015) REVENUES (%)

MARKET SHARE (ACCESSES)

COMPANY OR GROUP

Gross

Net

Fixed tel.

Mobile phones

Broadband

Pay TV

Vivo/GVT

32.1%

31.1%

34.3%

29.3%

29.1%

9.2%

America Móvil

25.3%

26.7%

25.8%

25.2%

31.6%

51.8%

Oi

20.5%

19.2%

35.5%

17.8%

25.8%

6.0%

Tim

12.4%

12.1%

1.3%

26.4%

-

-

Sky

4.9%

5.8%

-

-

-

28.8%

Nextel

2.9%

2.9%

-

0.74%

-

-

Other

1.9%

2.2%

3.1%

0.6%

12.7%

4.1%

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

Total Brazil

Source: . Accessed: January 28th, 2016.

Market domination by some application and content companies has re‑ sulted in other harmful impacts on development of the Internet, especially in emerging markets. Facebook, for example, generates such a concentration of data traffic that some users mistake the Internet for the application itself. An article published in 2015 in Quartz reveals that millions of Facebook users have no idea that they are on the Internet (Figure 7.2).

10 See Chapter 1 of this book for a comparative analysis of the prices of fixed and mobile broadband in several countries including the prices collected by the ITU.


177

FIGURE 7.2. Percentage of respondents who agree with the statement that “Facebook is the Internet” Nigeria Nigéria

65%

Indonesia Indonésia

61%

India Índia

58%

Brazil Brasil USA EUA

55%

5%

Source: Mirani (2016).

These data reveal an economic problem and not just mere semantic con‑ fusion. Considering that there are tens of millions of users in Brazil subject to the practice of zero rating and accessing the major application platforms such as Facebook, we will have decisive effects on the evolution of the Internet as a whole. Why? Startups, civil society organizations, publishers, governments and all agents who work in the Internet market will have to enter into con‑ tracts with these big economic groups to communicate on a large scale. That is, the economic power of these large groups can influence the incentives and decisions of various market players and civil society. Therefore, these eco‑ nomic groups may restrict the formation of the consciousness of millions of people according to their commercial interests, as a result of the concentra‑ tion of the information market. Not to mention the aspects of security and privacy protection, because the five most accessed application companies in Brazil and in the world participate in the massive and arbitrary surveillance system conducted by the US National Security Agency (NSA).

Social impacts of zero rating The concentration of millions of users on the same platforms and social networks influences the development of social relations and ends up affecting the shape and economic structure of the society, which is the objective basis

178

flávia lefèvre

on which to build the legal and political systems which, in time, condition the forms of social consciousness. That is, “the mode of production of material life conditions the process of social, political and spiritual life in general. It is not a man’s consciousness that determines his being, but, on the contrary, it is his social being that determines his consciousness” (MARX, 2008, preface). Therefore, these aspects must be at the center of disputes now in progress regarding the right to neutrality. In this context, the practice of zero rating added to limited data plans has potential to harm a country’s cultural, educa‑ tional and economic development, as it reduces the ability of citizens to freely form their consciences and to exercise their right to free flow of information. These practices also restrict business innovation and the dissemination of new digital services and content. Programs like Internet.org by Facebook (marketed under the name of Free Basics) and based on the practice of zero rating are being implement‑ ed in developing countries (17 by January 2016, including India, Colombia, Ghana, Kenya and Zambia). Although they claim to promote digital inclusion in partnership with governments and mobile operators, they actually pose threats to democratization in the Internet. The logic of this program is, as disclosed, providing connection and free access to poor people for a certain time, certain applications and content, whose suppliers are associated with Facebook. The objective is to make these users interested in the Internet so that in the future they start paying for it.11 Thus, the first of the threats is that governments, upon joining with Facebook, get comfortable with the idea that universal Internet access can hap‑ pen via mobile network. Worse than that, in a system not guided by the public interest, but by the economic interest of the companies participating in this project. That is, consumers of the most remote areas of the country would be subject to an inadequate navigation, different from what the Internet is: free, unrestricted and open. Pedro Soares Ramos, a net neutrality expert, says that the existence of zero rating plans has an adverse impact on com‑ petition. They can also be harmful to users, as they contribute to replicate a division between full Internet access for those who can pay for it and limited access to some applications for the neediest populations. (SOARES RAMOS, 2015, p. 9)

11 Available on: . Accessed: January 28th, 2016.


179

Internet.org involves creating a discriminated class of users; those who do not have income or infrastructure available for full access will be on the margin of net neutrality. Furthermore, the program creates unwanted effects in the environment of competitive and innovative Internet. Recently, regula‑ tory authorities in India – one of the first countries where Internet.org was deployed – concluded that zero rating plans bring about negative effects and, therefore, must be limited, and programs like Facebook’s Internet.org must be combated (BARGAWA, 2015). This argument was accepted by the Telecom Regulatory Authority of India (TRAI), which decided to ban the practice of dif‑ ferential pricing for data plans based on content (PURAKAYASTHA, 2016; GOEL; ISAAC, 2016). In November 2015, the Federal Public Prosecutors Office (Ministério Público) issued Technical Note 02/2015, signed by the Cyber Crimes, Media and Technology working groups, indicating a series of illegalities related to Internet.org, including the potential damage to net neutrality gurantees.12

Consumer rights, continuity, neutrality and zero rating Zero rating also impacts basic consumer rights, especially the right to choose and the prohibition of tie-in sales. Considering the market power of the leading providers of application and content, it is their economic inter‑ ests that will determine what will no longer be deducted from the franchise and not the consumers’ interests and needs. So when hiring access to the Internet, consumers who cannot afford an unlimited plan will be subject to the plan offered by the provider. When the franchise runs out, they will only have access to a restricted part of the Internet universe. This restriction of Internet access is contrary to the inherent guarantees of essential public services – especially continuity, as we saw above, since when data limits are reached the speed of provision may be reduced, but there can be no blockage of access. Unrestricted access to the Internet must be main‑ tained, taking into account the definition of the MCI, “the system consists of the set of logical protocols, structured worldwide for public and unrestricted use, in order to enable data communication between terminals through dif‑ ferent networks.”

12 Available on: . Accessed: April 20th, 2016.

180

flávia lefèvre

Thus, contrary to what has been argued by some, the obligation to main‑ tain the continuity of service provision and not break the neutrality by block‑ ing and discriminating is not abrogated just because the contract between the service provider and the consumer informs that the access will be restricted to certain content or applications. The information provided by the supplier to the consumer regarding an illegal practice does not change the fact that it is an illegal practice.

Zero rating and public interest The practice of zero rating, as discussed above, is not harmful per se. It could be used to serve the public interest if governments and public entities for‑ mulated contracts with providers to ensure that access to public services were not deducted from the data limits. For example, there would be no charge when the user accessed websites to file income taxes, pay taxes, file police reports, use public health services, enroll in social programs, participate in public consul‑ tations etc. That is, access would be sponsored by the government, in order to expand the reach of public services and the exercise of citizenship.

Monitoring Finally, it is essential that the presidential decree that will be issued to regulate the MCI establish which agency will be in charge of the supervision of the practices adopted by the Internet service providers. The draft decree released in January 2016 and submitted to public consul‑ tation by the Ministry of Justice indicates that the entity in charge of investi‑ gating violations of neutrality will be Anatel, which should do it according to the guidelines set by CGI.br. This seems right to us, especially in light of the provisions of article 24, I and II, MCI. Therefore, we have to define what topics are related to the provision of Internet access and to telecommunications, and thus subject to the regulato‑ ry power of Anatel, and what topics are not considered telecom-related and therefore should be dealt with according to the CDC and subject to the regu‑ latory power of CGI.br.

CONCLUSIONS We hope, then, that the government will take the necessary legal mea‑ sures, with the support of legal instruments available both in the LGT and in


181

the MCI, not only to universalize broadband infrastructure, but also to ensure access to Internet services. In this sense, the regulation of net neutrality will play a key role so that business practices like zero rating will not jeopardize the development of the market, with anticompetitive behaviors that affect not only the guarantee of access to the essential service of Internet access, but also to fundamental rights such as the free flow of information, freedom of expression and privacy.

REFERENCES BANDEIRA DE MELLO, C. A. Serviço Público e sua Feição Constitucional no Brasil. In: Direito do Estado: novos rumos. Tomo 2 – Direito Administrativo. São Paulo: Editora Max Limonad, 2001. BARGAWA, AK et al. Net Neutrality: DoT Committee Report. New Delhi, [sd]. Available on: . Accessed: December 22nd, 2015. DECLARATION OF SÃO PAULO – result of NET Mundial April 2014. Available on: . Accessed: February 12th, 2016. DEMICHEL, A. Le droit administratif – essai de reflexion théorique. Paris: Librairie Général de Droit et de Jurisprudence, 1978. GALPERIN, H.; MARISCAL, J.; BARRANTES, R. The Internet and Poverty: opening the black box – Estudo realizado pela Diálogo Regional sobre a Sociedade de Informação. Victoria: IDRC/CRD3, 2014. Available on:. Accessed: Dec. 22nd, 2015. GETSCHKO, D. NETmundial: um gol de placa. PoliTICs, Rio de Janeiro, Instituto Nupef, v. 18, p. 1214, Aug. 2014. GOEL, V.; ISAAC M. Facebook Loses the Battle in India Over Its Free Basics Program. New York Times Feb 8th, 2016. Available on: ;. Accessed: Feb. 12th, 2016. GROTTI, D. A. M. O Serviço Público e a Constituição Brasileira de 1988. São Paulo: Malheiros, 2003. JUSTEN FILHO, M. Teoria Geral dos Serviços Públicos. São Paulo: Dialética, 2003. ______. Course of Administrative Law. São Paulo: Saraiva, 2005.

182

flávia lefèvre

LEVY, E. Marco Civil da Internet: a visão dos provedores de acesso fixo emóvel. Audiência Pública, Comissão Especial – PL 2126/11. Brasília, Sinditelebrasil, 12 jun. 2012. Available on: Audiência. Accessed: Feb. 12th, 2016. MARCUS, J. S. Network Neutrality Revisited: Challenges and Responses in the EU and in the US”, a study on behalf of the European Parliament’s IMCO Committee IP/A/ IMCO/201402, PE 518.751(2014). Available on: Accessed: July 30th, 2015. MARX, K. Contribution to the Critique of Political Economy. Online English translation available on: . Accessed: October 29th, 2016. MEDAUAR, O. Direito Administrativo Moderno. 2. ed. São Paulo: Revista dos Tribunais, 1997. MEIRELLES, H. L. Direito Administrativo Brasileiro.17. ed. São Paulo: Malheiros, 1992. MINISTRY OF COMMUNICATIONS OF INDIA. Net Neutrality Dot Commitee Report, May 2015. MIRANI, L. Millions of Facebook users have no idea they’re using the internet. Quartz, Feb 9th, 2015. Available on: . Accessed: Jul. 30th, 2015. PURKAYASTHA, P. Internet power to the people. The Hindu, Feb. 10th, 2016. Available on: . Accessed: Feb. 12th, 2016. ROSSINI, C.; MOORE, T. Exploring Zero Rating Challenges: Views From Five Countries. A Public Knowledge Working Paper, jul. 2015. Available on: . Accessed: Jul. 30th, 2015. SIMET. Available on: . Accessed: April 19th, 2016. SOARES RAMOS, P. H. Zero rating: uma introdução ao debate. PoliTICs, Rio de Janeiro, Instituto Nupef, v. 21, p. 210, Aug. 2015. Available on: . Accessed: October 29, 2016.

|8|

TAXATION OF BROADBAND SERVICES Peter Knight [ Fernand Braudel Institute of World Economics ]

T

INTRODUCTION

he next few years will bring substantial challenges to Brazil and to the telecommunications industry. These challenges have to do with taxation, regulatory framework and incentives for investment. The already high tax burden on the industry, including broadband services, increased in 2016 because of the rise of the Tax on Goods and Services (Imposto sobre Circulação de Mercadorias e Prestação de Serviços – ICMS) in 11 states and in the Federal District and the end of exemption from two fed‑ eral taxes on smartphones, tablets, routers, notebooks, and other computer goods. The ICMS is by far the most important tax on telecommunications services, including broadband (fixed and mobile). Even though the tax in‑ crease will improve the fiscal situation of governments (federal or state, in the case of the ICMS) in the short run, a high tax burden has a negative im‑ pact on inflation and economic growth. These tax increases were imposed in early 2016, a time of economic crisis, with unemployment and inflation rising and the GDP shrinking. Brazil has experienced a rapid expansion of telephony and broadband penetration, as documented in several chapters of this book. From the first privatization initiatives in 1998 through 2014, operators invested 310.7 billion BRL in the expansion, modernization and improvement of service quality. Moreover, between 1998 and 2014, operators spent 44.7 billion BRL to ac‑ quire concessions for the provision of services, and the government collected approximately 22.4 billion BRL with privatization initiatives (TELEBRASIL, 2015, p. 8). The result of these investments – including private and govern‑ ment investments (with funds from concessions and privatization) – was the

184

peter knight

creation of one of the world’s largest telecommunications infrastructures, with a total of 346 million subscriptions to all services at the end of 2015.1 The Brazilian telecommunications sector, in addition to generating about 60.1 billion BRL in taxes every year, plays an important role in the development of the country, employing 498,000 people and producing rev‑ enues equivalent to 4.2% of the Gross Domestic Product (GDP), in spite of all the challenges it faces. Of the total value produced by the telecom industry from 2002 to 2014, governments have appropriated 59% in taxes (TELEBRASIL, 2015b). Very little of that amount, paid by the citizens, was actually used to support the telecom industry and promote digital inclusion, as we will demonstrate in this chapter. The tax burden, which is already one of the world’s heaviest, accounted for over 43% of the industry’s net revenues in 2015. It rose to 46% in 2016, without including the collection of sector funds, as we will show below. In Brazil, there is still a large contingent of people who do not have Internet access. Most of them do not have enough income to purchase broadband ser‑ vices and/or live in remote areas. According to the annual survey of the Center of Studies on Information and Communication Technologies (Cetic.br) 2014, in Brazil, only 55% of the population over 10 years old accessed the Internet in the three months prior to survey.2 Moreover, 76% of users accessed the web through a mobile phone or smartphone (CGI.BR, 2015, p. 335, 353 and 354, tab. C2 and C16), almost the same percentage of those who accessed the Internet using a computer or tablet. This suggests that many users used both mobile and fixed broadband. In 2015, smartphones outnumbered computers, and in 2014 the number of computers sold fell for the first time in 30 years, a decrease of 10%.3 In 2014, smartphone sales reached 55.2 million, a 57% increase in comparison

1

Includes mobile phones, landlines, fixed broadband and pay TV. Available on: . Accessed: February 29th, 2016.

2

For comparison, data published by the World Bank shows that in 2014 the percentage of the population who accessed the Internet at least once was 57.6% in Brazil, 44.4% in Mexico, 64.7% in Argentina, 72.4% in Chile, 61.5% in Uruguay, 70.5% in Russia, 84.3% in South Korea and 86.4% in the United States. Available on: . Accessed: September 11th, 2016.

3 Available on: . Accessed: March 5th, 2016.

taxation of broadband services

185

with 2013. In 2015, however, it dropped to 47.8 million, a decrease of 7.4 mil‑ lion smartphones or almost 9%.4 In 2016, smartphones, which are highly prized and important tools for so‑ cial inclusion – became more expensive due to the increase in the foreign ex‑ change rate (many of the smartphones’ components are still imported) and in taxes. Law 13,241 of December 30, 20155 withdrew the exemption of PIS/PASEP and Cofins (two federal tributes) on smartphones, tablets, routers, notebooks, and other computer goods. This exemption was part of the digital inclusion program of the federal government. As mentioned above, broadband services, both fixed and mobile, cost more because of the increase in the ICMS. In this chapter, the high tax burden, which is passed along to consumers in the prices they pay, and the negative impact of excessive taxes on digital in‑ clusion will be analyzed. This impact is even greater because of the diversion to other uses of the vast majority of the collection from three specific taxes in the sector. The chapter concludes with recommendations on how to reduce this tax burden, promoting digital inclusion and the economic, social and po‑ litical development of the country.

TAXATION OF THE TELECOM INDUSTRY In 2005, the information service industry (mainly telecommunications) was already the most taxed among the eight major sectors of the Brazilian economy. Indirect taxes alone6 amounted to 57.2% of all industry revenues, compared to 42.8% for electricity, gas and water and 24.7% for manufactures. Indirect taxes on information services were also much higher than indirect taxes in other industry groups, including “financial services”, “transportation” and “other services” (Figure 8.1).

4

Available on: . Accessed: October 29th, 2016.

5 Available on: . Accessed: March 1st, 2016. 6

An indirect tax is a tax on transactions of goods and services. The tax base is the purchase and sale price.

186

peter knight

FIGURE 8.1. Indirect taxes on consumption in Brazil, 2005 Indirect taxation on consumption, 2005 Net tax on consumption

Taxes

100% 80% 60% 40% 20%

Real estate

Other services

Transportation, storage and distribution

Financial brokering, insurance and pension funds

Agriculture and livestock raising

Manufactures

Electricity, gas and water

Information services

0%

Source: Werneck (2008, p. 23).

Such taxation (this includes taxes, fees and “contributions” in Brazilian terminology) is a legacy of the time when telephony was a service consumed mostly by the upper and upper-middle classes and businesses. Back then, Telebras was a federal monopoly and its subsidiaries were state monopolies. At that time, it was difficult to tax the income of individuals. But it was easy to use the Telebras system of monopoly state-owned telephone companies in each state to collect federal, state and municipal taxes via telephone bills. And this has also been the case for utilities that have been natural mo‑ nopolies such as electricity and gas. In general, Brazil overtaxes consump‑ tion while having poor performance in collecting traditional property taxes. For example, the Urban Building and Land Tax (Imposto Territorial Urbano – IPTU) collects less than the Tax on Motor Vehicles (Imposto sobre Veículos Automotores – IPVA) and the Rural Property Tax (Imposto sobre a Propriedade Territorial Rural – ITR) has been notoriously ineffective.


187

Taxes are concentrated on sectors with fewer taxpayers and high econom‑ ic activity. The stronger the monopoly or oligopoly, the higher the taxation, because that was easier to monitor and collect.7 The situation in the second decade of the 21st century is very different for three reasons: 1. With the end of the monopoly of state-owned companies in the Telebras system, competition among telecommunications companies and Internet service providers in the private sector tends to reduce the prices. 2. Because of the expansion of telephony (fixed and mobile) and Internet – the result of massive investments by private sector companies – in 2014, 93% of Brazilian households had at least one telephone, fixed or mobile (IBGE, 2015, tab. 1960), and 55% of the population over 10 years of age accessed the Internet in the three months prior to the annual survey of Cetic.br (CGI.BR, 2015, p. 335, tab. C2). Therefore, the high taxation is now levied on almost the entire population, not just on the upper classes. 3. The taxation system has become much more efficient. All formal busi‑ nesses in any industry have to issue electronic invoices that go to the tax authorities the moment they are issued. Therefore, these transactions are visible to the tax authorities and applicable taxes are collected automat‑ ically. Consequently, the tax authorities no longer need to prioritize the collection of taxes from a few large companies, as they did in the past. Moreover, the high tax burden imposed on telecommunications is con‑ trary to modern taxation principles. One of these principles is to tax all sectors of the economy at a low uniform level in terms of added value taxes (such as the ICMS), thus avoiding distortions that prevent the efficient allocation of investments and consumption. This principle provides for a few exceptions, such as a zero rate for essential goods or goods manufactured by small com‑ panies. The ideal strategy for a tax reform should include reducing tax rates, broadening the base and simplifying administration (IBRD, 2003; IMF, 2011). In the 21st century, the telecommunications industry, for reasons detailed in Chapter 1 of this book, is a strategic sector. The taxation of this sector is the 7

José Roberto Afonso, a taxation expert, notes that telecommunications must be seen in this context. These are services that everyone needs to use but they also have the heaviest tax burden. It is difficult to change the taxation on operators without making an overall change in this philosophy, that is, the objective should be to create a new, simpler and more efficient taxation system (correspondence via e-mail, March 8th, 2016).

188

peter knight

highest among 18 countries selected for comparison purposes. These coun‑ tries were selected because of their size and relevance to the telecommunica‑ tions sector in the world. They accounted for 55.4% of the world’s population in October 2015 (Table 8.1).

TABLE 8.1. Taxes levied on the telecommunications sector in 18 countries, October 2015 countries

taxes on:

Gross revenue

Net revenue

Brazil

30.2

43.2

Argentina

20.7

26.2

Portugal

18.7

23.0

Italy

18.0

22.0

Colombia

17.4

21.0

Peru

16.7

20.0

France

16.4

19.6

Chile

16.0

19.0

Russia

15.3

18.0

United Kingdom

14.9

17.5

Mexico

13.8

16.0

Spain

13.8

16.0

USA

11.8

13.4

India

10.9

12.2

Australia

9.1

10.0

South Korea

9.1

10.0

Japan

7.4

8.0

China

2.9

3.0

Source: Teleco (2015, Annex 3, p. 11).

The exceptionally high taxation of this sector is not only a distortion that reduces consumption and investment efficiency. It is also contrary to the digital inclusion policies promoted by federal, state and local governments, as high tax‑ es result in higher prices for broadband and other telecommunication services. Table 8.2 shows the extensive list of federal, state and local taxes that bur‑ den the telecommunications industry and, consequently, broadband services.


189

Table 8.2. Detailed taxes on the telecommunications sector in 2015 Acronym

Name

Calculation basis

Rate

IRPJ

Corporate income tax

Actual profit

15% additional 10%

CSLL

Social contribution on profit

Result of the year, before provision for IR

9%

PIS

Social Integration Program

Gross revenue

0.65%

Cofins

Contribution to Social Security Financing

Gross revenue

3%

State

ICMS

Tax on the circulation of goods and services

Gross revenue

25 to 35%

Municipal

ISS

Tax on services

Service price

2 to 5%

Fust

Fund for universalization of telecommunications services

Gross revenue

1%

Funttel

Fund for the technological development of telecommunications

Gross revenue

0.5%

TFI – Installation inspection fee

26 BRL to 48,000 BRL

TFF – Operation inspection fee

50% of TFI

Federal

Fistel

Funds and fees

Telecommunications inspection fund

Condecine

Contribution for national movie industry development

12% of TFI

CFRP

Contribution to the promotion of public broadcasting services

5% of TFI

PPDESS

Price for the right to deliver telecommunications services and for the right for satellite operation

Telecommunications services listed in Attachments from Resolution

Fixed value

PPDUR

Public price for the right to use radio frequencies

Application of formula defined in the Resolution

Formula

Source: Levy (2015).

The total of these taxes in 2014 was 60 billion BRL, or 43% of the opera‑ tors’ net income. On average, the taxes accounted for 59% of the added value produced by operators from 2002 to 2014 (TELEBRASIL, 2015b).

peter knight

190

Tax on the circulation of goods and services (ICMS) By far, the highest tax is the ICMS, levied by the states and the Federal District. In 2015, it ranged from 25 to 35% of the gross revenue of the opera‑ tors, depending on the state. At the end of 2015, 11 states and the Federal District decided to raise the tax, instead of lowering it (Table 8.3).

TABLE 8.3. Increased ICMS rates in 2016 States

From

To

From

AL

27%

30%

Jan/11/2016

AP

25%

29%

Feb/1/2016

CE

27%

30%

Feb/25/2016

DF

25%

28%

Jan/1/2016

MG

25%

27%

Jan/1/2016

PE

28%

30%

Jan/1/2016

RN

27%

30%

Jan/28/2016

RS

25%

30%

Jan/1/2016

SE

27%

30%

Jan/1/2016

TO

25%

29%

Jan/1/2016

a

RJ

29%

30%

Mar/28/2016

ROa

35%

37%

2016

Note: a It includes 4% for the state fund to tackle poverty in RJ. In RO we included 2% of the same kind of fund. Source: Teleco (2016).

Sector funds Three federal funds are statutorily destined to support the telecommu‑ nications sector: (i) the Fund for Universalization of Telecommunication Services (Fundo de Universalização dos Serviços de Telecomunicações – Fust); (ii) the Telecommunications Inspection Fund (Fundo de Fiscalização das Telecomunicações – Fistel); and (iii) the Fund for the Technological Development of Telecommunications (Fundo de Desenvolvimento Tecnológico das Telecomunicações – Funttel). The General Telecommunications Law (Lei Geral de Telecomunicações – LGT) of 1997 created the Fust, which was established by Law 9,998 in 2000. The Fust is funded primarily by the 1% contribution on the gross revenues of


191

telecom operators and by part of the Fistel revenues. Fistel was created in 1966 by Law 5070 to finance the inspection of telecommunications services. Fistel revenues come from two inspection fees: installation (TFI) and operation (TFF) of phone numbers, 50% of grant revenues from concessions, grants, radio use permits and those arising from fines provided for in the LGT. The main source of revenues of Funttel is a 0.5% contribution on the gross revenues of telecom companies, resulting from the provision of tele‑ communications services. This does not include canceled sales, discounts, ICMS, PIS and Cofins.8 From 2001 to September 2015, almost 89 billion BRL were collected in current prices (not adjusted for inflation) for these funds (Table 8.4).

TABLE 8.4. Collection of federal funds to support telecommunications (brl millions in current prices) Period

Funttel

Fust

Fistel

Total

127

1,045

4,430

5,602

2002

203

1,100

2,694

3,997

2003

214

531

1,288

2,032

2001

2004

242

711

1,978

2,932

2005

272

595

2,055

2,922

2006

302

634

2,019

2,955

2007

335

1,166

3,084

4,585

2008

377

1,444

6,417

8,238

2009

392

1,436

4,910

6,738

2010

421

987

3,431

4,839

2011

494

2,718

7,268

10,480

2012

542

1,925

4,919

7,386

2013

545

1,713

4,914

7,172 11,120

2014

591

1,757

8,772

2015

633

3,049

5,359

9,042

Total

5,690

20,811

63,357

90,039

Source: Telebrasil (2015b, tab. 10.1, p 151). Update for the full year 2015 by Telebrasil via email.

An audit of the Federal Audit Court (Tribunal de Contas da União – TCU) verified to what extent Fust and Fistel were used as the source of financing they were supposed to be. According to the rapporteur of the case, Minister 8

For details, see Ferreira (2015).

192

peter knight

Bruno Dantas: “data clearly demonstrates the gap between the amount col‑ lected and the application of funds in the main objective that motivated the creation of Fistel, i.e., to improve the inspection of telecommunications ser‑ vices.” The rapporteur also commented that “the differences show that the data published by Anatel – responsible for annually elaborating the account‑ ability of Fistel and Fust accounts – does not accurately portray the real situa‑ tion of these funds” (TCU, 2016). Among the uses of these funds, TCU highlighted the payment of meal vouchers for civil servants, investment in the North-South and in the EastWest railways, modernization and revitalization of aircraft, salaries and pen‑ sions of civil servants, and rural and urban social security benefits. Of all funds collected for Fust from 2001 to June 2015, Minister Dantas informed that only 1.2% were used for promoting universal access to telecommunica‑ tions services. Furthermore, of the nearly 100 billion BRL collected for Fust since 2001 and for Fistel since 1997, just over 6% were allocated to the uses for which they were created, presenting “a portrait of the chaotic condition of planning in Brazil” (TCU, 2016). Despite several attempts in Congress to use Fust to promote digital inclu‑ sion, starting with a bill (PL 1,481/07), by February 2016 this and other bills with the same goal had not passed yet. As for Funttel, whose goal is “to en‑ courage the process of technological innovation, promote the training of hu‑ man resources, promote the creation of jobs and promote access of small and medium companies to capital resources, so as to increase the competitiveness of the Brazilian telecommunications industry”,9 only 31% of the funds raised were allocated for this purpose (KERN, 2015). Why has the vast majority of sector funds been confiscated by the Ministry of Finance and not allocated to universalize telecommunications services? Why not ensure the necessary resources to foster innovation, labor force training and industry competitiveness? The only answers can be the voracity of federal taxation, aimed at reducing the public sector deficit, the low priority of the afore‑ mentioned objectives and the lack of planning (and transparency), as indicated by TCU.

9

Available on: . Accessed: Feb 1st, 2016.


193

In addition to the considerations above, the Ministry of Communications (MC) has evidence, based on several studies on price elasticity of demand,10 that a reduction in current rates on broadband could actually increase tax collection at all levels of the federation. This is because the increase in the demand for broadband services would offset the reduction in prices. The re‑ duction would also promote stronger economic growth thanks to the increase in broadband penetration. The MC analyzed three studies conducted over the period 2008-2010 that support this hypothesis regarding price elasticity of demand. In 2012, the min‑ istry made its own estimates, based on 2011 data, from the household survey by Cetic.br.11 Estimates ranged from 1.0 to 3.36. MC’s own study estimated a value of 1.4. The ministry then proposed that the four federal taxes on broadband services (PIS, Cofins, Fust and Funttel) were reduced to zero and that the state ICMS rates were reduced from an average 26.7% to a uniform rate of 10%. These reductions would result in a 24% drop in prices, assuming that all reductions would be passed on to end consumers. If these exemptions had been put in practice in 2012, the MC estimates – using a conservative estimate of 1.5 for the price elasticity of demand – that about 12 million additional broadband connections would have been pur‑ chased by 2016. This would have resulted in a loss of ICMS revenue of only 1.7% that year. This estimate does not take into account the increased eco‑ nomic growth that would accompany the increase in the use of broadband. Studies by the World Bank estimate that a 10% increase in broadband pene‑ tration in developing countries is associated with an increase of 1.4% in the growth rate of per capita gross national income. If the increase in the growth rate had been included in the estimate of the MC, tax cuts would lead to an increase in revenues from ICMS in the states. In short, the taxation on the telecommunications industry in Brazil is one of the highest in the world. It is also the highest among the eight major sec‑ tors of the Brazilian economy. This tax goes against fundamental principles of tax theory and discourages digital inclusion. And finally, most of the money

10 The price elasticity of demand is the ratio of the percentage change in quantity demanded of a good or service and the percentage change in its price. A negative value higher than one (1) indicates that a reduction in price increases the demand more than proportionally. 11 Correspondence with Pedro Lucas da C. P. Araújo, Project Manager, Department of Broadband, Telecommunications Secretariat, June 28th, 2013.

194

peter knight

raised by the three funds that should support the development of the sector has been systematically diverted for other purposes.

TAX POLICY RECOMMENDATIONS It is essential to reduce the high tax burden on the telecommunications in‑ dustry in general and broadband in particular to promote digital inclusion and economic, social and political development in Brazil. A gradual reduction of taxes, both federal and state, could prevent a sharp fall in short-term revenue. The two measures together would tend to increase tax revenues in the medium term. As for the reduction of the specific tax burden on the industry, sector rates (Fistel, Fust and Funttel) should be reduced so as to ensure the balance be‑ tween the amounts collected and effectively invested, according to their re‑ spective purposes. The collection of these funds should not be used for pur‑ poses that have nothing to do with the telecommunications industry. In the case of Fust, in accordance with the existing legislation, resources cannot be used for universal services on a private basis, like broadband, but only under the public regime, like telephony. Using the resources of Fust in order to promote universal broadband will require that the Congress pass Draft Law 1,481/2007 or the like, or change the model of provision of telecommunica‑ tions services established in the LGT from 1997. As detailed in Chapter 3 of this book, these issues were under discussion inside and outside the federal government in February 2016. Entities in the telecommunications industry, like Telebrasil, play an active role in promoting the political and regulatory changes they believe promote the development of the telecom sector.12 On the tax issue, specifically, Telebrasil and its members have been saying for years that the high tax burden is an obstacle to the development of the sector and the digital inclusion of the population.

12 The organizations formed by medium and large operators and their suppliers, especially the National Union of Telephone Companies and Mobile and Personal Service (Sinditelebrasil) and the Brazilian Association of Competitive Telecommunications Service Providers (Telcomp) are key members of an umbrella organization, the Brazilian Telecommunications Association (Telebrasil). In February 2016, Telebrasil had 48 members: large and medium-sized telecommunications companies, equipment and software suppliers, consulting firms and professional associations (11 business associations and industry associations). Therefore, it is quite representative in the sector.


195

The opinions of these important organizations about the high tax burden on the industry and the misuse of sector funds largely matches the analysis and recommendations presented in this chapter (TELEBRASIL, 2015a).

CONCLUSIONS Reducing the prices charged for telecommunications services in Brazil, especially for broadband (see Chapter 1, Tables 1.4 and 1.5), should be a na‑ tional priority. Lower prices are necessary to make broadband more affordable to the less favored classes, especially in rural and remote areas. The reduction of the tax burden should have a favorable impact on both prices and invest‑ ments by operators. Brazilian governments (federal, state and municipal) should recognize the essentiality of telecommunications and quality broadband for all citizens, adopting public policies that encourage the digital inclusion of more and more Brazilians. Reducing the tax burden on the telecommunications sector should be part of a broader strategy to use ICT to promote the country’s development.

REFERENCES AFONSO, J.R. Interview with author, 8 March 8th, 2016. IBRD, RM; ZOLT IN Introduction to Tax Policy Design and Development. Draft prepared for the course on Practical Issues of Tax Policy in Developing Countries. Washington, DC.: World Bank, April 28 to May 1, 2003. Available on: . Accessed: March 1st, 2016. COMITÊ GESTOR DA INTERNET NO BRASIL (CGI.br). Survey on the use of information and communication technologies in Brazilian Households – ICT Households 2014. São Paulo: CGI.br, 2015. Available on: . Accessed: January 23rd, 2016. FEBRATEL; TELEBRASIL; FENAINFO. Projeto Brasil Digital Inovador e Competitivo 2015-2022: Contribuições para Formuladores de Políticas Públicas. São Paulo: LCA Consultores, 2014. Available on: . Accessed: February 3rd, 2016. FERREIRA, M. A arrecadação e a destinação dos fundos das Telecomunicações. Apresentação para Audiência Pública. Brasília, Ministério das Comunicações. June 17th, 2015. Available on: . Accessed: September 11, 2016. INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA (IBGE). Pesquisa Nacional por Amostra de Domicílios 2014. Rio de Janeiro: IBGE, 2015. Available on: . Accessed: December 7th, 2015. INTERNATIONAL MONETARY FUND (IMF). Revenue Mobilization in Developing Countries. Washington, DC: Fiscal Affairs Department 8 March 2011. Available on: . Accessed: March 1st, 2016. KERN, S. S. Serviços de Telefonia no Brasil 2015. Apresentação na Câmara dos Deputados, Comissão de Fiscalização, Financeira e Controle, December 3, 2015. Available on: . Accessed: February 3rd, 2016. LEVY, E. Telecomunicações no Brasil. Apresentação na Audiência Pública na Comissão de Fiscalização Financeira e Controle da Câmara dos Deputados, September 2015. Available on: . Accessed: January 23rd, 2016. MANCINI, S. Governo sanciona MP que aumenta tributação de bebidas e eletrônicos. In: O Estado de S. Paulo, January 1st. 2016. Available on: . Accessed: March 1st, 2016. TELEBRASIL. 59o Painel Telebrasil 2015: Carta de Brasília 2015. August 2015a. Available on: . Accessed: February 3rd, 2016. TELEBRASIL. O Desempenho do Setor de Telecomunicações no Brasil: séries Temporais 9M15. Available on: . Accessed: February 3rd, 2016. TELECO. O Desempenho do Setor de Telecomunicações no Brasil Séries Temporais 1S16. Available on: . Accessed: October 29th, 2016. ______. Cargas Tributárias. January 22nd, 2016. Available on: . Accessed: February 8th, 2016.


197

TRIBUNAL DE CONTAS DA (TCU). TCU apresenta relatório sistêmico do setor de telecomunicações. January 2016. Available on: . Accessed: February 8th, 2016. Werneck, R. Tax reform in Brazil: an evaluation at the crossroads. Texto para discussão n. 558, Departamento de Economia, PUC-Rio, Jan. 2008. Available on: . Accessed: December 5th, 2015.

INSTITUtions

|9|

THE RNP NETWORK AND NEW PARTNERSHIPS Eduardo Grizendi Michael Stanton [ Brazilian Education and Research Network – RNP ]

T

INTRODUCTION

he Brazilian Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP) is today a Social Organization (Organização Social – OS), recognized in 2002 by the then Ministry of Science and Technology (Ministério de Ciência e Tecnologia – MCT).1 Its history had begun earlier, as an MCT program, launched in 1989 to provide Brazil with a nation‑ al academic network, in response to demands from the academic community for access to digital means of communication. These demands were first ex‑ pressed in the mid 1980s, but became stronger after 1987, certainly because of the creation of the NSFNET network2 – based on Internet technology (TCP/ IP) – by the National Science Foundation (NSF) of the United States.3 A key aspect has characterized the concept of an academic network: it ag‑ gregates traffic from the networks of all member institutions, allowing com‑ munication between them without using other network services provided by other entities, for example, by commercial carriers. Therefore, the academ‑ ic community manages its own interconnection network. This aggregation of demand for connectivity enables economies of scale and consequent cost

1

Successively renamed Ministry of Science, Technology, and Innovation (MCTI) in 2011, and Ministry of Science, Technology, Innovations and Telecommunications (MCTIC) in 2016.

2

The National Science Foundation Network (NSFNET) was an Internet financing program, sponsored by the National Science Foundation (NSF) between 1985 and 1995, to promote a network of education and research in the United States. Available on: . Accessed: January 22nd, 2016.

3

The origins of the academic network in Brazil are retold in an excellent dissertation, written in Portuguese, by Marcelo Savio Carvalho, especially in Chapter 4 (CARVALHO, 2006).

202

eduardo grizendi • michael stanton

reduction, allowing the community to provide its individual users and other academic entities with privileged access conditions. This chapter traces the evolution of the RNP network, from its begin‑ ning to the present day, and summarizes the vision we have for its future development. The chapter reports RNP network’s contribution to the expansion of broadband Internet in Brazil and describes the partnership between RNP and other participating institutions. The evolution of the network and of RNP it‑ self, as a government project in the early years, and as a non-profit civil society since 1999, has progressed in parallel. We will discuss both here. This account is necessarily summarized because of space limitations. Fortunately, there are additional sources that can be consulted, particularly Stanton (2010), which features a history of RNP between 1987 and 2010, in English, and the RNP website, in particular “Our history”,4 with maps of the backbone network at each stage of its evolution.

OVERVIEW OF THE GROWTH OF THE RNP NETWORK Back in 1988, construction of a simpler network with Bitnet5 technology (Because It’s Time Network) had already begun, providing (non-interactive) services to users of e-mail and file transfer – two of the first applications also available on the Internet. This cooperative network grew by voluntary acces‑ sion of new nodes, each of which paid the cost of a link to an existing node. The ability of these links ranged between 1.2 and 9.6 Kbps (kilobits per second). By 1991, the Brazilian Bitnet already reached 42 institutions in all regions of the country, from the National Laboratory for Scientific Computing (LNCC) in Rio de Janeiro and the São Paulo State Research Foundation (Fapesp), in São Paulo city, as illustrated in Figure 9.1. These two network nodes main‑ tained international connections to the United States, with capacities between 4.8 and 9.6 Kbps.

4


5


THE RNP NETWORK AND NEW PARTNERSHIPS

203

FIGURE 9.1. Bitnet network topology in Brazil in 1991

Source: Stanton (1993)

Unlike the Bitnet network, of a cooperative nature, the initial objective of the RNP program was to build a planned and robust network to meet the needs of research institutions across the country, including major universi‑ ties. The technology chosen for this network was Internet (TCP/IP), which was already spreading around the world – a movement strongly influenced by the example of NSFNET. Compared to Bitnet, Internet technology had three major advantages: (1) the Internet allows the use of any technology support‑ ing unreliable transport of digital data packets, such as Wi-Fi or wired local networks; (2) the Internet enables the use of personal computers and, more recently, tablets and smartphones, as network terminals, each one configured and dedicated entirely to its user; and (3) the Internet offers support to an end‑ less variety of alternative applications, including interactive ones. The flexibil‑ ity and extensibility of Internet technology has allowed its tremendous growth to the present day. The first RNP network, then the called the National Research Network (Rede Nacional de Pesquisa) was deployed in 1992 to give access to the cam‑ pus networks of connected institutions through a hierarchically structured

204


interconnection network, with a single point of presence (PoP)6 in each con‑ nected state as well as in the Federal District. The PoP concentrated traf‑ fic from the campus networks of client institutions in the state (or Federal District), where each client had one or more connections to the PoP. The eleven original PoPs were interconnected by the national backbone, which was a mesh (multiply connected) network of interstate links with capacity of 9.6 and 64 Kbps, as illustrated in Figure 9.2. This network had a single 64 Kbps international connection from the PoP of São Paulo to the United States. Many of the nearly fifty client institutions had already been client nodes of the earlier Bitnet network.

FIGURE 9.2. RNP backbone from 1992 with connections of up to 64 Kbps

Source: RNP

6

From the term “Point of Presence”.


205

More than twenty years later, in 2015 the RNP network was offering access to more than 1,200 campi across the country, through the backbone network illustrated in Figure 9.3, with an exponential increase in the capacity of the backbone and of the international connections. Figure 9.4 illustrates the in‑ crease between 1992 and 2014. In this picture, the top three graphs show the evolution (measured in Kbps) of (1) the maximum capacity limit of Ethernet technology, from 10 Mbps in 1992 to 100 Gbps in 2014; (2) the aggregate capacity of international communication, ranging from 64 Kbps in 1992 to 25 Gbps in 2014; and (3) the average capacity of the links of the national RNP backbone, varying from 35 Kbps in 1992 to 12 Gbps in 2014. Figure 9.4 also indicates the timeline divided into six phases, discussed in the next section.

FIGURE 9.3. RNP’s backbone in 2015 with links of up to 20 Gbps

Source: RNP

206


FIGURE 9.4. The evolution of transmission capacity (Kbps) of RNP networks between 1992 and 2014

Source: RNP

DEVELOPMENT PHASES OF RNP AND ITS NETWORKS In the period from 1992 to the present, RNP has evolved from a project to an organization. Its activities have grown greatly in order to meet the de‑ mands presented by those ministries of the federal government which have sustained RNP. At the same time, RNP has been offering services qualita‑ tively different from those of alternative, usually commercial, Internet service providers, especially after the start of commercial Internet services in Brazil in 1995. Since RNP was already operating the only Internet network within the country before the start of commercial services, it also participated initially in the spread of the Internet in Brazil, acting as an upstream provider of nation‑ al and international connectivity for new commercial Internet providers that emerged in this period (see Chapter 10). However, this activity came to a halt


207

in 1999, when the attention of RNP once again focused on providing services to research and education institutions, taking the name National Education and Research Network (Rede Nacional de Ensino e Pesquisa), with the retention of the original acronym, RNP. 1999 was the year when RNP became an institution. Before then it was an MCT project, with unique features, such as the maintenance of a large technical and administrative staff, individually remunerated through maintenance grants, and the direct payment by the Ministry of backbone telecommunications contracts. This institutionalization sought a solution compatible with federal Law 9637/1998, the Law of Social Organizations (OS), a name given to private non-profit entities engaged by a government agency to carry out public interest activities, through a long-term manage‑ ment contract. So in 1999, the National Education and Research Network Association (AsRNP) was formally created, whose members were the self‑ same participants of the RNP project. Additionally, the Inter-Ministerial RNP Program was created, involving MCT and the Ministry of Education (MEC), and it was agreed to sign a management contract with AsRNP, in order to provide connectivity and other related services to their research units, universities and institutes. The considerable growth of the RNP network is mainly due to the par‑ ticipation of MEC, which sustains financially the federal universities (UFs) and federal institutes of education, science and technology (IFs), which con‑ stitute the majority of RNP’s customers. At the same time, MCT began the process of recognizing AsRNP as an OS, and this was completed in 2002. From 2002 on, the conditions of the Law of Social Organizations were ap‑ plicable, and the duration of the management contracts with RNP could thereafter extend to five years. These changes had significant consequences for the stability of the RNP initiative, with the possibility of medium and long term prospects, allowing planning, including of its technological evolution, with important conse‑ quences for the capacity and robustness of the networks operated by RNP. We will address this matter in further detail ahead. The growth of the RNP networks was continuous and it can be defined in six stages, distinguished from one another by changes in the capacity and/or the technology used in the backbone. These phases also create a timeline for RNP, enabling us to point out relevant events in its development (Table 9.1).

208


TABLE 9.1. The six phases of RNP’s backbone until 2015 Phase and dates Phase 1: 1992-1994

Backbone

Important Events

Links up to 64 Kbps.

RNP operates the first Internet network in the country with interstate connections.

Extension to 15 PoPs. Connections using telephone circuits.

Phase 2: 1995-1999

Links up to 2 Mbps. National coverage with 27 PoPs. 2 to 5 international connections at 2 Mbps.

Start of the commercial Internet in the country, using RNP’s and commercial backbones (1995) High-Speed Metropolitan Networks Project (Redes Metropolitanas de Alta Velocidade – ReMAV) for experimental use (1997-present, in some cases). Foundation of the RNP Association and the Interministerial RNP Program (MEC-MCT) (1999)

ATM and Frame Relay Technologies.

45 Mbps connection to Internet2 (Ampath Project with Florida International University (FIU): 2001-2004).

Nodes up to 155 Mbps, connections up to 28 Mbps.

RNP next generation workshop (2001).

International Connections at 155 Mbps and 45 Mbps

Project GIGA – interstate optical fiber testbed network (2003-7).

Phase 4: 2004

SDH and PDH.

RedClara begins operation.

Connections up to 622 Mbps.

Commoditya Internet access contracted in Brazil

Phase 5: 2005-2010

Optical links at 2.5 and 10 Gbps reaching 10 capitais

Redecomep: building out owned optical metro networks in larger cities (2005-).

Phase 3: 2000-2003

RNP qualified as a Social Organization (2002).

10 Gbps connections to the US in partnership with São Paulo state network (ANSP) and FIU (2005); Use of commercial Internet Exchange point (IXP)b in Miami for commodity traffic (2009-). Phase 6: 2011-

Optical Backbone with links at 3 e 10 Gbps provided by Oi (2011-)

Future Internet Brazilian Environment for Experimentation (FIBRE) (2011-)

40 optical metro networks (2014).

Veredas Novas (New Paths): connections to non-capital cities (2012-). 2 Container-based data centers (CDCs) (2013-). 20 Gbps connection to Miami (2013); 110 Gbps to Miami (2016)

With the commercialization of the Internet from 1989 on, there has been a separation between the academic Internet, where only academic institutions can communicate through academic networks, and the commodity Internet for other users. To allow communication between academic institutions and commercial entities, there are connections between the academic Internet and the commodity Internet that carry traffic also called commodity. In the case of international traffic, access to the commodity Internet may be provided within the country through the PoP of an international provider, or it may be provided abroad, reached through an international connection of the academic network, as has been the case of RNP before 2004 and after 2009.

a

b Internet Exchange Point: place where there is exchange of traffic between two or more networks (see Chapter 10)


209

Since this book is about broadband, we should consult the relevant in‑ formation in Table 9.1 and Figure 9.4. We notice that in the early years of RNP (phases 1 and 2), one could not say that the RNP network offered broad‑ band service. At the time, the capacity of connections between RNP’s PoPs and university campi rarely reached 1 Mbps. In fact, they were much lower most of the time. There were exceptions, of course, but these were limited to metropolitan areas such as the High-Speed Metropolitan Networks (Redes Metropolitanas de Alta Velocidade – ReMAVs), and were considered experi‑ mental. We might associate the beginning of broadband use only to phase 3, after the institutionalization of RNP, when the capacity of the backbone began to increase significantly. Between 1999 and 2005, the average capacity of the backbone increased from 1 Mbps to 2 Gbps, and the RNP network began to incorporate the traits that characterize it today, with large backbone capacity to meet the needs of the access networks for the campi being served – a true broadband network. Figure 9.4 shows in 2011, for the first time in the history of RNP, the con‑ vergence of the three indicators of capacity (limit of technology, international connectivity and backbone) around the value of 10 Gbps. This “balanced” sit‑ uation was the result of several factors that guided the evolution of RNP since the creation of its first network in 1992. These processes are described in the following section.

THE EVOLUTION OF RNP’S NETWORKS In 2000, Phase 3 of the RNP backbone was launched. It was called RNP2. This was the first outward sign of its reorganization as a non-profit company on its way to becoming an OS, which began in 1999. In April 2001, RNP held a workshop on the next generation of the RNP network with the participation of scientific users and technologists, where the future of RNP networks was discussed.7 The main concerns expressed were with the improvement of its quality: higher capacity, lower latency, higher availability, increased security, access to high-performance computing and storage facilities, inclusion of new academic institutions and access arrange‑ ments between the RNP network and the networks of commercial providers

7

A more complete report of the workshop held in 2001 (in Portuguese) may be found in Stanton (2001).

210


in Brazil, which reflected the demand for user access to the academic network from their homes. As potential routes to be explored by RNP, the following directions were indicated: • Building metropolitan optical networks with its own infrastructure to interconnect all the institutions of interest in each city with capacities of at least 100 Mbps; • Increasing the national backbone capacity to offer end-to-end connec‑ tions of at least 100 Mbps between users in different cities (the limit in 2001 was 10 Mbps among 12 of the PoPs, and much lower among the rest); • Using smaller towns with metropolitan networks as connection points to other isolated and neighboring institutions (rather than directly to the PoP in that state); • Creating Internet exchange (peering) points (IXPs) to allow increased interconnection capacity between commercial providers, and between them and RNP; • Creating an experimental testbed for long-distance optical networks to study and validate equipment, as well as to provide the means to develop and validate high-performance applications. Some of these directions were chosen based on experience already ac‑ quired during the previous backbone phases, especially ReMAV (since 1997), which had brought the first contact with new network technologies, and with capacities very much higher than those available in the RNP production network. The ReMAVs were isolated from the backbone and considered as testbeds. The choice of most directions chosen was based on the example given by the evolution of academic networks in other countries, mainly in Europe, the US and Canada, with which there had already been an exchange of information and experience, not only by network specialists, but also by researchers and educators increasingly aware of the impact of networks in these countries. Next we will describe the technological and organizational evolution of RNP’s network after 2001. Instead of addressing this issue through a sin‑ gle chronology, we prefer to deal separately with the evolution in parallel of


211

different aspects of this evolution, each at its own speed. The issues to be addressed are: • Connectivity hierarchy. • Infrastructure sharing. • The potential of optical communication – ReMAV and Giga. • Owned optical infrastructure – Redecomep. • Owned optical infrastructure – economies of scale and swapping of rights of use. • The search for partnerships. • Expanding the network to areas of difficult access. In the discussion of each of these aspects we refer to different moments in the evolution of RNP that contributed to its development, with ramifications for the advancement of broadband networks.

Connectivity hierarchy From the point of view of the RNP network, the basic unit served is a cam‑ pus network (or that of a single building), which typically consists of a set of local area networks, which may be wired, using Ethernet (IEEE 802.3), or wire‑ less, using Wi-Fi (IEEE 802.11), or a combination of both of these. RNP pro‑ vides external connectivity for this campus network. This connection is made to another campus of the same institution in the same city, or to a point of ag‑ gregation (PoA)8 or to the nearest point of presence (PoP)9 of RNP’s backbone. Typically, this connection uses either a telecommunications circuit leased by RNP or RNP’s own infrastructure, typically a metropolitan optical network.10 In a city housing a RNP backbone PoP, the metropolitan network gives direct access to it. In other cities with a metropolitan network, the network has a PoA, which connects – usually via a leased telecommunications circuit – to the PoP in the same state. In January 2016, there were approximately 400 campi served by metropolitan networks, and 900 campi with a leased inter‑ city link directly to the nearest PoP. In most cases, these are terrestrial links, 8

Point of Aggregation.

9

Point of Presence.

10 The Redecomep program (http://www.redecomep.rnp.br) has built more than 40 metropolitan optical networks in 26 capitals and other cities where there are multiple campus networks.

212


but in some parts of the country there is no land-based telecommunications infrastructure, so the alternative is satellite infrastructure, despite its high cost and reduced capacity.

Infrastructure sharing The Inter-Ministerial RNP Program is maintained by the federal govern‑ ment to provide network services to (mostly federal) university campi spread around the country. To meet this challenge, the Program seeks to provide appropriate communications infrastructure, and the costs are covered by the management contract. The choice of RNP as the single connectivity provider for all campi, rather than the individual leasing of connections from commer‑ cial Internet service providers, allows for large economies of scale. Today the RNP network has more than 1,300 locations served throughout the country. Since it operates a national backbone with a PoP in each state, RNP is able to provide network service to all of them, using this backbone, complemented with access connections between the campi and the local PoP. The national backbone has been the most visible component of this shared infrastructure, but it has always been complemented by international connec‑ tions to ensure access to the global Internet. Another opportunity to explore economies of scale through infrastructure sharing occurs when there are several campi in the same city, usually a state capital, and home to a backbone PoP, typically, but not always, located on the campus of the main federal university in the state. This subject will be explored later in the section “owned optical infrastructure”.

The potential of optical communication – ReMAV and Giga In the late 1990s, in different parts of the world, long-distance optical trans‑ mission systems, used both in terrestrial networks and submarine cables, began to adopt the use of Dense Wavelength Division Multiplexing (DWDM), which greatly multiplied the capacity of each fiber pair. Brazil also benefited from this enhancement: in addition to the arrival in the country of three new international submarine cable systems that would use this technology,11 Embratel adopted it in its land routes and several initiatives involving the deployment of optical cables and lighting up optical fiber. Eletronet was created as the result of an initiative of electricity transmission companies associated with state-owned Eletrobras and a 11 See Chapter 14.


213

foreign partner.12 In 2000, Eletronet completed the construction of an optical net‑ work reaching 16 states, mainly by way of OPGW (Optical Ground Wire) technol‑ ogy, used in high-voltage electrical power lines.13 The Eletronet network adopted DWDM to share its optical fibers among different clients. This widespread and growing adoption of optical transmission technology had already been noticed by RNP in 1997, when it launched the ReMAV initia‑ tive to build metropolitan optical networks using borrowed fibers.14 Academic partners in a city could use their own equipment, provided by RNP, to create a private multiple-access network, with capacities available in the range from 34 to 155 Mbps. We should note that these capacities were far higher than the RNP backbone capacity, which would only surpass 2Mbps in 2000, with the start of phase 3 of the backbone, during which backbone connections would increase up to 40 Mbps. ReMAV demonstrated the advantages of sharing optical infrastructure, using its own equipment and avoiding the leasing of commercial connection services. However, it depended on access to optical fibers. In ReMAV, all fi‑ bers used were loaned temporarily to the project, free of charge. We notice that, in general, the ReMAV were experimental networks and were not con‑ nected to the RNP backbone. For production use, in the same period, the Universidade Federal Fluminense (UFF) designed and deployed its own optical fiber infrastructure, that spanned 12 Among other DWDM initiatives, we should mention the Geodex company, which deployed an optical route in the South along the ALL railroad; Intelig, which deployed an optical route in the Southeast and Northeast, along the FCA and CFN railways, and was later acquired by TIM; Metrored, which deployed the Santos – São Paulo – Rio de Janeiro – Belo Horizonte route, along the MRS railway, and metropolitan networks in São Paulo, Rio de Janeiro and Belo Horizonte, and was later acquired by Brasil Telecom and subsequently absorbed by Telemar, now called Oi; Pegasus, which deployed optical routes in the state of São Paulo and the São Paulo – Curitiba – Florianópolis – Porto Alegre optical route, along highways, and later acquired by Telemar, now Oi; Engeredes, which deployed routes in the Southeast, especially the Campinas – Brasília route, using the Petrobras oil pipeline, and later absorbed by its parent company, Algar Telecom; Netstream, which deployed metropolitan networks in São Paulo, Rio de Janeiro and Belo Horizonte, and later acquired first by AT&T and then by Embratel; and Impsat, which deployed routes and metropolitan networks in the capitals of the Southeast and South, later acquired by Global Crossing, which in turn was acquired by Level3. 13 A map of the Eletronet network is available on: . Accessed: January 22nd, 2016. 14 A brief description is available on: . Accessed: January 22nd, 2016.

214


the center and some adjacent neighborhoods of Niterói, RJ.15 This initiative followed the general line of the ReMAV, with the significant difference that UFF became the sole owner of the optical infrastructure for the entire service life of the cables. This network initially made it possible to interconnect 12 UFF campus networks in Niterói, with 155 and 622 Mbps connections. This set of networks jointly enjoyed access to the RNP backbone. Thus, even before the April 2001 workshop, the potential benefits of op‑ tical technologies were already recognized at RNP. One of the participants of this workshop was a senior researcher in optical communications from the Center for Research and Development in Telecommunications (CPqD), located in Campinas, SP. CPqD was already developing optical transmission technology with DWDM, and it would soon create a “spin-off” company called Padtec,16 to manufacture and market optical products. The interaction be‑ tween RNP and CPqD deepened after this workshop, and in late 2002 Project Giga – a joint proposal to set up a testbed to validate and develop the use of optical networks – received funding of more than 50 million BRL from the Fund for the Technological Development of Telecommunications (Funttel), through the Financier of Studies and Projects (Finep). This project enabled CPqD to specialize in setting up and operating a DWDM optical testbed net‑ work providing connectivity between participating laboratories, with data rates of 1 and 10 Gbps, and the development of certain categories of user ap‑ plications, while RNP would monitor the development of high performance applications between these laboratories. This joint project lasted from 2003 to 2007, and involved researchers from almost 50 institutions from much of the country. The testbed was 750 km long, reaching the cities of Campinas, São Paulo, São José dos Campos and Cachoeira Paulista, in São Paulo state, and Rio de Janeiro, Niterói and Petrópolis, in the state of Rio de Janeiro, provid‑ ing optical connections of 1 and 10 Gbps to research laboratories in some 20 institutions (Figure 9.5). The optical fibers used were provided without cost by Telefonica (now Vivo), Intelig (now TIM), Telemar (now Oi) and Embratel. The optical equipment was supplied by Padtec.17

15 Available on: . Accessed: February 20th, 2016. 16 Available on: . Accessed: February 20th, 2016. 17 A description of the project is available on: . Accessed: February 20th, 2016. See also Scarabucci (2005), Barros (2005).


215

FIGURE 9.5. Project Giga testbed network: (a) geographical map, (b) topology of the optical connections to participating institutions

Source: Barros (2005) Acronym Explanation UNICAMP University of Campinas LNLS National Synchrotron Light Source Laboratory (Campinas) CPqD ICT Research Center (Campinas) Net Ops GIGA Network Operations Center (Campinas and Rio de Janeiro) CAS GIGA backbone node in Campinas SPO GIGA backbone node in São Paulo LAB T Telefonica Laboratory (São Paulo) INCOR Heart Institute (São Paulo) USP University of São Paulo (São Paulo) SJC GIGA b+B23 backbone node in São José dos Campos INPE National Institute for Space Science (São José dos Campos) CTA Air Force Aerospace Technology Center (São José dos Campos) CPTEC Center for Weather Forecasting and Climate Studies/INPE (Cachoeira Paulista) RJO GIGA backbone node in Rio de Janeiro

216


UERJ Rio de Janeiro State University (Rio de Janeiro) TELEMAR LEME Telemar exchange, in Leme, Rio de Janeiro PUC Catholic University (Rio de Janeiro) IMPA Mathematics Institute (Rio de Janeiro) CBPF Brazilian Center for Research in Physics (Rio de Janeiro) IME Military Institute for Engineering (Rio de Janeiro) UFF Fluminense Federal University (Niterói) FIOCRUZ Oswaldo Cruz Foundation (Rio de Janeiro) CRT – EBT Embratel Laboratory (Rio de Janeiro) NCE UFRJ ICT center of Federal University of Rio de Janeiro (Rio de Janeiro) LNCC National Scientific Computing Laboratory (Petrópolis)

For RNP, there were several benefits in having taken part in this collabo‑ ration with CPqD. The main ones were: • Access for the first time to a long-distance network with transmission rates of 1 to 10 Gbps. This prepared it to plan, set up and operate fu‑ ture networks with these capabilities. • Allowing some users to develop and use advanced applications in high-performance networks, including international collaborations, from 2004 on.18 Two years after starting the Giga project, RNP deployed phase 5 of its backbone, based on the leasing of optical channel infrastructure of 2.5 and 10 Gbps, connecting 10 PoPs. Costs were similar to circuits with Synchronous Digital Hierarchy (SDH) technology – of much lower capacity – used in the previous year (phase 4). Figure 9.4 clearly shows the impact of this innova‑ tion, which made the average transmission capacity in the backbone increase by a factor of more than 1,000 between 1999 and 2005. This new RNP net‑ work received the name of Ipê network in 2005.

Owned optical infrastructure – Redecomep Until the mid-2000s, access to the local PoP of the RNP backbone posed the biggest challenge for sharing connections from campus networks. Access 18 In November 2004, one of the first uses of the new RedClara 622 Mbps link between São Paulo and Europe (described below) was the 400 Mbps international transfer of data by physicists between the Rio de Janeiro State University (UERJ) and the Supercomputing 2004 Conference in the USA. In order to have access to São Paulo, they used a 1 Gbps connection through the Project Giga network. This ease of exchanging large volumes of data quickly became extremely useful for the effective participation of Brazilian scientists in international collaborative projects with intensive exchange of data. By 2009, such exchanges occurred at rates close to the physical limits of the 10 Gbps connections made available in that year.


217

was usually implemented through a commercially provided point to point cir‑ cuit for each campus. Starting in 1997, the ReMAVs and the UFF network in Niterói had shown how to build a metropolitan optical network in a dozen cities. But at the time there was no way to connect these metropolitan networks with each other with capacities similar to those used locally. However, as we just saw, in 2005 RNP started to provide connections of 2.5 and 10 Gbps between ten capitals. How could RNP use the Ipê network to enable high capacity collaboration between institutions in different cities? Clearly a new model of metropolitan network was needed for these cities. In 2004, a feasibility study was carried out in Belém, capital of the state of Pará, by RNP and the local federal university (UFPA), of a optical connectivity solution for a set of campi of research and higher education institutions located in Belém and a neighboring city, Ananindeua. The novelty was to extend the owned infrastructure model adopted by the UFF network in Niteroi to serve all 32 campi of the 12 institutions to be connected in Pará. Nine of them were public institutions and three were private universities. This study concluded that the return on investment of building an owned infrastructure of aerial op‑ tical cables, as well as the acquisition of network equipment, would be offset in approximately two years, due to the savings in the cost of leasing connectivity services (STANTON, 2005). As a result of the great interest aroused by this study, in the second half of 2004 MCT approved funding for two metropolitan network construction projects: Metropolitan Network of Belém (MetroBel), to serve Belém and Ananindeua, and Community Networks for Education and Research (Redecomep), to serve the other 26 cities in the country where RNP operated its PoPs. The funds covered project costs and the construction of these metropolitan networks. The operation, maintenance and upgrade of these new networks would be the responsibility of their local administrators, usually the institutions served by the metropolitan network. In 2005, RNP started construction of these owned metropolitan networks to provide access between each PoP and the campi of the user institutions in the same city, through the Redecomep program.19 The Redecomeps are high capacity networks deployed by RNP in metropolitan areas served by the PoPs of the RNP backbone, and in some other cities with two or more user insti‑ tutions. These networks allowed the provision of high capacity access to the 19 Available on: . Accessed: January 22nd, 2016.

218


campi, typically at 1 or 10 Gbps, using Ethernet technology, usually in a ring configuration, to provide redundancy. The model consists of the deployment of an owned optical infrastructure and the setting up of consortia of participating institutions in order to ensure their self-sustainability. At the end of 2014, these metropolitan networks were already in operation in about 40 cities, including capitals and important upstate cities. By 2016, 400 campi participating in the Redecomep networks were thereby connect‑ ed to the Ipê network backbone. Figure 9.6 shows the location of these networks by state and Table 9.2 identifies the cities where networks were built.

FIGURE 9.6. The distribution of Redecomeps by state at the end of 2014

Source: RNP.


219

TABLE 9.2. Region, city and name of Redecomeps in 2014 Southeast region ES MG

RJ

Center-West Region

Vitória

MetroVix

DF

Brasília

GigaCandanga

Belo Horizonte

RedecomepBH

GO

Goiânia

MetroGyn

Ouro Preto & Mariana

Network Inconf.Edu

MS Campo Grande

Rede

Niterói

MetroNit

MT

Pantaneira

North Region

Petrópolis Rio de Janeiro

RedeRio

Campinas SP

São Carlos

MetroSanca

São Paulo

MetroSampa

AC

Rio Branco

RB MetroNet

AM Manaus

MetroMAO

AP

Macapá

MetroAP

Belém

MetroBel

Região Nordeste AL

Cuiabá

Altamira

Maceió

RAAVE

Salvador

Remessa

Marabá

Juazeiro and Petrolina

RedeVASF

Santarém

Fortaleza

GigaFor

RR

Boa Vista

RedeBV

TO

Palmas

MetroTins

Campina Grande

MetroCG

South Region

João Pessoa

Rede ICTIPB

PR

Curitiba

PE

Recife

Rede Ícone

SC

Florianópolis

Remep-FLN

PI

Teresina

Rede Poti

RS

Porto Alegre

MetroPOA

RN

Natal

GigaNatal

SE

Aracajú

MetrAju

BA CE

MA São Luís PB

PA

Castanhal

Owned optical infrastructure – economies of scale and swapping of rights of use The construction of an optical network infrastructure brings both benefits and duties. The duties include the effort (and cost) to maintain and, when necessary, repair this infrastructure, especially after cable cuts, which affect aerial installations (on poles or electrical transmission py‑ lons), as well as underground (ducts) or underwater installations. The owner of the infrastructure needs to ensure its integrity throughout its lifetime. The benefits include the ability greatly to enhance the (present and future) capacity to transmit information in three different ways: (a) overprovisioning the cable to be installed, acquiring a larger number of

220


fibers than initially required; (b) increasing the transmission rate in the fibers, through upgrade of the transmission equipment (transponders and detectors at both ends of the fiber link); and (c) multiplexing the transmit‑ ted signal channels in a single optical fiber using DWDM techniques, as mentioned above. Methods (b) and (c), which leverage the existing optical fiber infrastructure, increasing its transmission capacity without physi‑ cally changing the optical cable, are sometimes referred to as the use of “scalable” infrastructure. Option (a) is always attractive, since the marginal cost of each additional fiber is a small fraction of the cost of the cable. The total installation cost is usually several times higher than the cost of the cable itself. Deploying spare fibers makes it possible to meet future demands. Option (b) allows exchanging transmission equipment for another with higher capacity. For example, in 2016 RNP was undertaking migration from 10 to 100 Gbps, due to the recent development of optical transmission technologies, without needing to change the optical fiber used. Option (c) takes advantage of the fact that the infrared radiation beams used to transmit optical signals can be generated using just a fraction of the spectral width (frequency band) available. In 2016, a single optical fiber using DWDM can transmit more than 100 optical channels simultaneously, each with a capacity of 100 Gbps of information, i.e., a total of 10 Tbps (terabits per second). This approach to network construction and management enables widespread sharing of optical fiber infrastructure, in which the unit of capacity to be shared could be a fiber pair or a fraction of the usable spectrum of a fiber pair. In other words, the owners of optical infrastructure can sell or lease a fraction of the avail‑ able capacity along a cable route, or they can swap this fraction for equivalent capacity in other infrastructure, along one or more other cable routes. Such swap agreements for long periods are very common between telecommunications op‑ erators, but this kind of arrangement is not limited to these operators. We may note the growth the installation of optical infrastructure belonging to organiza‑ tions, whose prime activity is not in telecommunications. Examples include elec‑ trical power companies, large corporations such as the oil company, Petrobras, state governments, or other network infrastructure owners, such as RNP. Thus, with the construction of its owned optical fiber infrastructure in metropolitan net‑ works, RNP acquired the clout to participate in swap agreements with other com‑ panies or institutions, in order to gain access to infrastructure in other locations


221

or routes. In practice, this may mean trading the use of fiber in metropolitan areas for fiber (or spectrum) in interurban fibers.20

The search for partnerships The experiences of Project Giga and the Redecomep program have shown that it is desirable, if not essential, for RNP to have partners in its optical network projects. Could this also be the case for long-distance connections, used in the construction of backbone and access connections to clients in upstate cities? In its search for alternatives to gain access to long-distance optical fi‑ ber resources, on several occasions, starting in 2001, RNP sought agree‑ ments with companies with long-distance fiber, mainly in the energy and telecommunications sectors. The first success came in 2003 in the collab‑ oration with CPqD in Project Giga, but the benefit did not extend to the production network. It was limited to groups of researchers from laborato‑ ries in around twenty institutions connected to the project network. This restriction was imposed by telecommunications companies that provided the fibers, since the RNP production network was already built using the same companies’ leased circuits. However, in order to use long-distance fiber optics in the operation of its backbone, RNP had to wait until 2009, when it signed a technical cooperation agreement with Oi, the Brazilian telecommunications compa‑ ny with the most extensive network at that time. The agreement provided for investments in research and development by Oi in order to fulfill the conditions laid down in “Research and Technological Innovation Project 001/10”, between RNP and Oi, under Consent Act 7828, of December 19th, 2008, of the federal telecommunications regulator, Anatel, related to the acquisition by Oi of Brasil Telecom. One of the annexes of this Consent Act, denominated “Conditions for the Consent Act, Item 9 – On Investment in Research and Development”, imposed on Oi the condition of signing an agreement with RNP, to support the provision of services and infrastructure of an ad‑ vanced education and research network in Brazil, through the ces‑ sion of transmission capacity in optical fibers for non-commercial use by the national research network (RNP), making feasible the 20 An example of these partnerships, both public-public and public-private, is discussed in Chapter 13 of this book – the Digital Beltway of Ceará.

222


national interconnection between universities and their campi, re‑ search centers, laboratories, teaching hospitals and museums, to generate knowledge and innovation through the academic network.

Following approval by RNP and Oi on April 30th, 2010, of the “Research and Technological Innovation – 001/10” project, joint work began for the gradual inclusion of optical infrastructure made available by Oi in the back‑ bone of the Ipê network. Oi made available the use of about 30 intercity op‑ tical channels from its own national optical infrastructure. This increased to 10 Gbps the connections between 15 PoPs, and to 3 Gbps the connections to nine other PoPs. Thus the federal capital and 23 of Brazil’s 26 states were now reached by multiple gigabit connections.21 Figure 9.7 illustrates this new backbone, which was rolled out in 2011. Subsequently, the agreement with Oi was amended in 2014, which is re‑ flected in the map of Figure 9.3, which increased to 21 the number of PoPs with 10 Gbps connexions, maintaining only three with 3 Gbps, and continu‑ ing to exclude the other three.22

21

The PoPs with 10 Gbps were in AL, BA, CE, DF, ES, MG, PB, PE, PR, RJ, RN, RS, SC, SE and SP. The PoPs with 3 Gbps were in AC, GO, MA, MS, MT, PA, PI, RO and TO. The capitals of the states of AM, AP and RR were not within the footprint of the optical infrastructure of Oi.

22 The new PoPs with 10 Gbps were GO, MA, MS, MT, PA and TO.


223

FIGURE 9.7. RNP’s backbone in 2011 with connections in multiple Gbps as a consequence of Anatel’s Consent Act

Source: RNP

This partnership with Oi enabled RNP to improve access to its clients, es‑ pecially those that participated in a Redecomep of the new gigabit PoPs. On the other hand, it created an almost total dependence on Oi’s optical infrastructure, and it was possible to identify some potential vulnerabilities caused by insuffi‑ cient redundancy, affecting the availability of the service offered by RNP. In 2010, the resurgence of Telebras as a telecommunications operator, largely responsible for carrying out the National Broadband Plan (PNBL), provided RNP with a new partner to help meet its challenges. This partner‑ ship primarily addresses services to customers in upstate cities, and will be commented on below. In addition to meeting the needs of these cities, the

224


partnership with Telebras made possible improvements in the availability of RNP’s backbone. Between 2012 and 2015, some additional backbone circuits were provi‑ sioned by Telebras: Brasília – Belém, at 1 Gbps, and 10 Gbps circuits on the following routes: Goiânia – Brasília, Brasília – Fortaleza, Fortaleza – Recife, Recife – Salvador and Salvador – Vitória, (see Figure 9.3). These circuits have increased the availability of PoPs connected to the North and Northeast rings of the Ipê network, providing protection to the traffic flows between the North/Northeast and Southeast/South, which had hitherto depended entirely on circuits from Oi.

Expanding the network to areas of difficult access The first map of the RNP backbone in 1992 (Figure 9.2) shows only 11 PoPs, leaving out 16 states. Only in 1999 was it possible to bring con‑ nectivity to all states, and the PoPs of six states in the North Region (AC, AM, AP, RO, RR, TO) used geosynchronous satellite connections, with low capacity and high latency. At that time there was simply no land-based telecommunications network in most of the Amazon region. However, the situation of telecommunications has evolved, especially since the pri‑ vatization of the state-run telecommunications monopoly in 1997-98, enabling the gradual replacement of satellite links by land connections over optical fiber, with much higher capacities. At the end of 2015, the backbone already reached the states of Pará and Tocantins with 10 Gbps, Acre and Rondônia with 3 Gbps, Amazonas and Amapá with 1 Gbps, and Roraima with 100 Mbps. Despite this progress, there are still RNP user institutions, especially in the Amazon region, linked via satellite to the Ipê network. Chapter 14 describes a new initiative – Amazonia Connected – to overcome some of these limitations by installing optical fiber in the river‑ beds of the Amazon region. This “digital divide” negatively affects telecommunications and network conditions not only in the Amazon region, but also in other remote areas far from the major population centers and the routes between them. However, with the growth of the geographical coverage by optical networks used in tele‑ communications, the number of connected users also grows.


225

VEREDAS NOVAS: A PROGRAM FOR CONNECTING UPSTATE USER CAMPI The access circuits for RNP user campi are mainly provisioned by three alternatives: • By RNP’s metropolitan optical networks (Redecomep). • By telecommunications service providers and operators. • By partners, through the Veredas Novas (New Paths) Initiative described below.

Veredas Novas In 2012, RNP launched a new initiative – Veredas Novas – with the aim of connecting, by 2014, all upstate RNP user institutions, with the following capabilities: secondary campi, at 100 Mbps; headquarter campi, at 1 Gbps. Veredas Novas is a joint initiative of the Ministries of Science, Technology, Innovations, and Communications (MCTIC) and Education (MEC) with the support of the National Association of Heads of Federal Institutions of Higher Education (Andifes) and the National Council of the Institutions of the Federal Network of Professional, Scientific and Technological Education (Conif). Technical collaboration to attain these targets for upstate campi, by provid‑ ing them with high-capacity connections, is currently based on partnerships with companies and organizations, amongst which, the following stand out: • Telecomunicações Brasileiras S.A. (Telebras); • ICT company of the State of Ceará (Etice); and • ICT company of the State of Pará (Prodepa). Unfortunately, the initiative failed to meet its goal for the end of 2014, which was to connect all upstate secondary campi with at least 100 Mbps and the headquarter campi with at least 1 Gbps. About 30% of the institutions were still served by circuits at less than 100 Mbps, including some served by satellite circuits between 2 and 18 Mbps. In any case, significant advances were made.

Veredas Novas – Telebras RNP and Telebras signed an agreement in 2010 for the implementation of access at 100 Mbps and 1 Gbps to the campi of user institutions in locations

226


with technical and commercial viability. By January 2016, 63 campi had been connected by Telebras, two at 1 Gbps and the others at 100 Mbps, using opti‑ cal fiber and radio.

Veredas Novas – Etice In 2012, RNP and Etice signed an agreement for the deployment of 100 Mbps and 1 Gbps access to all upstate user organizations in the state of Ceará. A pilot project was carried out and connected the campi of Crato, Iguatu and Aracati through 100 Mbps links, in addition to the campus of the University for the International Integration of Afro-Brazilian Speakers of Portuguese (Unilab) in Redenção, at 1 Gbps. By January 2016, 34 campi of user institu‑ tions had been connected at capacities of 32 to 100 Mbps, one at 500 Mbps and one at 1 Gbps. This partnership between RNP and Etice is emblematic for RNP and has become a reference model for partnerships with other state IT and ICT companies.23

“Pre-Veredas Novas”: a model for the participation of local and regional telecommunications service providers In 2013, RNP realized it would be necessary to seek alternatives to the Veredas Novas program to meet some critical needs. Therefore, it outlined a new strategy to connect the upstate campi of user institutions. The Veredas Novas initiative had not found any market response to provide at least 100 Mbps connections, except in Ceará, through the successful partnership be‑ tween RNP and Etice. This new strategy, dubbed “Pre-Veredas Novas”, intro‑ duced a new type of contract, using new Terms of Reference with initially less stringent service conditions and the following characteristics: • Circuits with minimum bandwidth of 20 Mbps. • A two year contract. • Scheduled, phased updates during the contract, up to 100 Mbps. • A preference for deterministic circuits. • If deterministic circuits unavailable, coverage through Metro Ethernet or MPLS circuits, using routed IP.

23 See Chapter 13 for more details on the evolution of the partnership model with Etice.


227

• Price ceiling for 100 Mbps fixed at 12,000 BRL per month, including taxes, for all regions, except the North. Along with the design of this new model, RNP focused on attracting lo‑ cal and regional service providers to participate in its bidding procedures for these campi, competing with the traditional operators who already supplied circuits to RNP. With this new strategy, it was possible to instigate and rely on the commitment of these service providers, bringing more competitivity to the contracting process. This resulted in circuits being leased with higher bandwidth and lower prices than those from traditional suppliers. With the traditional suppliers, RNP has found it difficult to increase the contracted ca‑ pacity and pay a fair price for it, whereas local and regional providers find it easier to accept this challenge. In order to attract these providers, RNP built a large contacts database by mining several existing databases provided by associations of Internet provid‑ ers. RNP’s participation in regional and national meetings of these providers became more extensive and enabled RNP to present its needs more clearly. The strategy had immediate effect. New providers emerged to participate in the process, presenting more aggressive bids than traditional suppliers, and, consequently, were selected and contracted. In February 2016, in addition to RNP’s traditional suppliers (Oi, Embratel, Telefonica, GVT, Algar Telecom and Intelig), about sixty local and regional providers provided circuits for the campi of user institutions. The number of these circuits is already higher and the quality of their service providers is often better than those of traditional suppliers.

INTERNATIONAL CONNECTIVITY Perhaps the main reasons for the use of networks by the research and education community are to access to information and knowledge available in other countries, as well as to discover effective opportunities for collabo‑ ration with foreign partners. Indeed, these were the driving forces that led the academic community in Brazil to seek participation in the first academic networks in the 1980s, leading to effective participation in Bitnet, from 1988 on, and the Internet, from 1992 on. In order to make network access and collaboration possible, it is essential to have international communication between the networks of the countries involved. Initially, the “other” country was the United States, and the first

228


international connection established had a capacity of 9.6 Kbps, allowing use of electronic mail and non-interactive access to documents. In 2015, Brazilian academic networks already had nearly 50 Gbps of ag‑ gregate capacity for international communication, including RNP and initia‑ tives of Brazilian states, which are sponsored networks such as projects for their respective foundations that support research: Academic Network at São Paulo – ANSP24 (São Paulo) by Fapesp and RedeRio de computadores25 (Rio de Janeiro) by Faperj. The graph in Figure 9.4 illustrates the evolution of the capacity used by RNP since 1992. There were significant increases associated with different phases of the backbone, as shown in Table 9.3. In 2016, this increased to 110 Gbps.

TABLE 9.3. Evolution of the international connectivity capacity of RNP Phase Start year Capacity

1

2

3

4

5

6

1992

1995

2000

2004

2005

2011

64 Kbps

2 to 10 Mbps

155 to 202 Mbps

300 Mbps

300 Mbps to 10 Gbps

25 Gbps

Source: RNP

The nature of the global Internet In order to understand the evolution of the international connections of RNP and the other Brazilian academic networks, ANSP and RedeRio, we must first understand how the Internet evolved in other countries, especially in the US. The Internet came into existence on January 1, 1983, interconnect‑ ing a collection of networks of the US government and academia. In 1985, the National Science Foundation (NSF), the leading US research funding agency, decided to set up a national network, NSFNET, adopting the Internet stan‑ dard.26 Over the next ten years, NSFNET was the core of the Internet, serv‑ ing not only as its backbone in the US, but also as a backbone for the global Internet, which began to spread to other countries and continents, reaching

24 Available on: . Accessed: January 27th, 2016. 25 Available on: . Accessed: January 27th, 2016. 26 Available on https://en.wikipedia.org/wiki/National_Science_Foundation_Network. Accessed: February 27th, 2016.


229

Brazil between 1991 and 1992.27 From 1992 to 1995, there were two Internet connections between Brazil and the US. The second connection was RedeRio, also with 64 Kbps. Both connections were to academic networks in the US. However, in 1989 the commercial use of the Internet had already begun in the US. The NSF decided to close down NSFNET in 1995, considering its mission of establishing a network to meet academic needs accomplished. In the future, academia would be served by commercial providers. For this reason, during phase 2 (1995-1999) of RNP’s backbone, its various (up to 5) international connections of 2 Mbps were made to networks of commercial providers in the US. In 2000, these connections were replaced with a single connection of 155 Mbps to an Internet exchange point (IXP) in New York, where RNP installed its own router.28 Throughout this period, RNP kept only one category of connection, first to the academic Internet, then to the com‑ mercial Internet, also called the commodity Internet. This situation changed again in 2001, with the creation of a new academic Internet in 1998, with the launch of the so-called Internet2, a new US aca‑ demic network maintained by a consortium of universities.29 Internet2 was based on a new national backbone to serve academic users only. In the fol‑ lowing years, Internet2 took over, for academic users of the global Internet, the role of global backbone, which had been occupied by NSFNET until 1995. In 2001, RNP installed a second connection to the US, at 45 Mbps, to Miami, where Florida International University (FIU) offered connectivity at low cost to the Internet2 network, through the Ampath project.30 With these two connections RNP began to separate its international traf‑ fic into two parts: academic traffic used the academic connection to Miami, and commodity traffic used the commodity connection to New York. This separation between academic traffic and commodity traffic still characterizes the Internet today. Thus we can speak of the academic Internet, consisting of academic networks in different countries, which are interconnected with each other by purely academic connections; and the commodity Internet, which 27 The first connection between Brazil and the Internet (9.6 Kbps) was deployed in 1991 between Fapesp and Fermilab, in Illinois. In 1992, it became RNP’s 64 Kbps international connection (Figure 9.2). 28 A router is a device used to interconnect different networks and route traffic between them. 29 Available on: . Accessed: February 28th, 2016. 30 The history of the Ampath project is available on: . Accessed: January 27th, 2016.

230


includes the networks of commercial providers, both national and interna‑ tional. However, the two Internets are interconnected in all countries where academic networks belong to the global academic Internet. The use of commodity transport meets the need for access to companies and government entities of all kinds, while academic transport is only for in‑ teraction through academic networks, such as RNP itself, i.e. for universities and research centers in other countries. By using academic transport, users of RNP can perform large transfers of digital content such as scientific data or very high resolution media. This type of application simply does not work using the connections between commercial providers because these networks lack the conditions to carry very intensive data streams. This subject is revis‑ ited in the next section. In 2004 (phase 4), RNP left Ampath to be a founding member of RedCLARA,31 a Latin American regional network of academic nature, with 14 national academic networks as partners, and an intercontinental connection, between São Paulo and the pan-European academic network (GÉANT). This initiative, initially funded largely by the European Commission,32 extended to Latin America its concern to create in Europe a hub for developing the aca‑ demic Internet outside of the US. In 2004, the commodity connection to the US was also replaced by com‑ modity connectivity provided through the PoP in Brazil of an international commercial provider. In 2005 (phase 5), RNP began to share with ANSP the connection that ANSP had established with FIU, partially funded by NSF, with a capacity that would increase from 622 Mbps to 2.5 Gbps by 2008.33 In 2009, this solution was replaced by the sharing with ANSP of two 10 Gbps connections between São Paulo and Miami (FIU), which were used to carry international traffic, both academic and commodity, where the previous use of an international commodity provider in Brazil was replaced by the participation of both ANSP and RNP in an IXP in Miami. In 2011 (phase 6), the capacity between São Paulo and Miami was doubled to four 10 Gbps connections, with the same 31 Available on: . Accessed: February 28th, 2016. 32 Available on: . Accessed: February 28th, 2016. 33 This collaboration has had several names since 2005: WHREN/LILA (2005-2009), Amlight (2010-2014) and Amlight ExP (since 2015). It received funding from NSF in successive editions of the International Research Network Connections program (IRNC). Available on: . Accessed: February 28th, 2016.


231

partners: ANSP and FIU. This collaboration with ANSP and FIU was refor‑ mulated and upgraded in 2016, increasing the total capacity to 220 Gbps. As for the academic connection from São Paulo to Europe via RedClara, this experienced some capacity increases, shared with other national networks in Latin America, to 2.5 Gbps in 2008 and to 5 Gbps in 2014.

A NEW MEANING OF “BROADBAND” The construction of metropolitan optical networks by RNP since 2005 raised the connectivity baseline of campi in these metropolitan areas, allow‑ ing the use of high-capacity end-to-end transmission by users in those loca‑ tions. In fact, these RNP metropolitan optical networks, when combined with capacity increases in the national backbone and international connections, have given a new meaning to the term broadband: nowadays it is possible to use network applications at the endpoint, that use the full capacity of the local network connection. In the process of expanding the capacity of its networks, RNP has discovered that there are users who can take advantage of that in‑ creased capacity, with processing-intensive applications or very high resolu‑ tion video and image transmission. An important article (DE LAAT, 2003) analyzed the characteristics of dif‑ ferent user applications of the new high-capacity networks that began to be used in the Netherlands around the year 2000. A graph taken from this arti‑ cle is reproduced in Figure 9.8, where three classes of user are depicted: A. “Domestic” users, who use e-mail, WWW and media streaming. Typically the data streams are short and require full routing for com‑ munication with arbitrary sites. B. “Corporate” users (corporations, universities, virtual organizations and laboratories), using the capacity of local area networks (100 Mbps to 1 Gbps), and making heavy use of VPNs (virtual private networks). C. Users of high-performance applications, with demands for data transmis‑ sion well over 1 Gbps, but usually to only a few previously determined des‑ tinations. Typical cases include scientific applications like radio-astrono‑ my, high-energy physics, databases in genetics and Earth observation. The article estimates that these three categories of users generate similar aggregate demands, i.e. ΣA≈ΣB≈ΣC. This would naturally have to be taken into consideration when designing a real network.

232


FIGURE 9.8. User classes according to bandwidth usage and type of connectivity

Source: De Laat (2003 p. 1003).

In practice, the big difference between the Internet networks of large-scale commercial providers and academic networks is that, in general, only the lat‑ ter have there category C users. For example, RNP networks today support only a few category C users, who may be supported by adopting high-capacity optical infrastructure. These users are typically scientific researchers dealing with large volumes of observed or computed data in areas such as Earth ob‑ servation, weather and climate, genetics, astrophysics and high-energy phys‑ ics, as mentioned above. To meet the needs of these users, it is essential that entire routes followed by their traffic have sufficient capacity to accommodate this traffic. By this we mean the entire chain: campus network – metropolitan network – backbone – international connection. The end-to-end capacity limit is the smallest link component along the datapath.

PHASE 7 OF RNP’S BACKBONE (2017) The evolution of RNP’s backbone, as seen above, moved from Kbps cir‑ cuits to Gbps circuits in about two decades. Throughout this evolution and at each stage, RNP has always stayed ahead of the market in the search for high


233

capacity links, often facing difficult negotiations with telecommunications providers because of the high cost of these circuits. On a number of occasions, these providers had no other customers for the capabilities that RNP sought, so pricing has always been exhaustively discussed between the parties. In fact, RNP’s demand for backbone capacity has always been high. It has often had to decide whether to lease connections from providers or to provision them using its own infrastructure. RNP’s needs accelerated between 2013 and 2015, as the number of campi served almost tripled. The number of applica‑ tions that require high capacities and minimum delay times also increased, so RNP needed increasingly high capacities in the backbone, in access networks to user institutions, in IXPs, and in international connections. The discussion of this ever-growing demand for a higher capacity back‑ bone led to a decision on strategic planning for 2020, to gradually replace the capacities currently contracted for the RNP backbone with capacities of its own scalable optical infrastructure. To acquire this infrastructure, it will be necessary either to build it, or to obtain the right of use of existing dark fiber, or even of optical spectrum (whole or fractional), in existing optical routes that connect the PoPs of RNP’s backbone. The first and important optical route which is the object of this new strat‑ egy is the route between Fortaleza and Porto Alegre. This choice is motivated not only by pressure for higher capacity routes in the Southeast ring (São Paulo – Rio de Janeiro – Belo Horizonte – Brasília), but also because of RNP’s participation in the BELLA Project (Building Europe Link to Latin America), described below, which plans to provide not one or two, but multiple wave‑ lengths of 100 Gbps, so that traffic can flow between academic networks in Europe and Latin America.

The BELLA Project BELLA aims at the direct integration of academic networks in Europe and Latin America, using a submarine cable being built by EllaLink, a joint ven‑ ture between Telebras and the Spanish company Islalink, which will connect Fortaleza to Portugal. The project also plans to deploy an terrestrial optical route Fortaleza – Porto Alegre – Buenos Aires – Santiago – Lima – Quito – Bogota (see Figure 9.9).

234


FIGURE 9.9. BELLA project

Source: RNP

In 2018, the EllaLink submarine cable will connect Brazil to the European continent without passing through North America, which nowadays func‑ tions as a “hub” of communication between South America and the rest of the world. The new cable will connect Sines, in Portugal, to Fortaleza, Ceará, with future intermediate accesses to Madeira, the Canary Islands and Cape Verde, and an extension to Santos (Figure 9.10). The cable will have four pairs of optical fibers, each with at least 10 Tbps transmission capacity. Forty percent of the spectrum of one of these fibers (at least 4 Tbps) will be for teaching and research activities between Latin American and European research networks, which comprise the BELLA consortium.


235

FIGURE 9.10. EllaLink Cable

Source: RNP

RNP is a member of the BELLA consortium, along with the Latin American regional research and education network, RedClara, and other European and South American research and education networks. The goal is to eliminate connectivity bottlenecks between Europe and Latin America, in order to facil‑ itate the exchange of data at high capacity, and to foster collaboration between researchers and educators from both regions. As the submarine cable con‑ nects only Portugal and Fortaleza at high capacity, it will be necessary to build access networks on both sides of the Atlantic Ocean, which will extend across the consortium’s member countries, to provide the benefits of increased com‑ munication capacity to partner research networks. Among the advantages of the BELLA design, we can highlight the re‑ duction of connection costs, low communications delay (latency) in large scientific experiments, and greater redundancy in case of problems on traditional routes. Moreover, these benefits will last throughout the ser‑ vice lifetime of the submarine cable, estimated at 25 years, and of the land infrastructure deployed.

236


The Fortaleza – Porto Alegre route So that BELLA can reach its goals, it will be necessary to invest in in‑ frastructure, both in the submarine cable and in the land network on both sides of the Atlantic. In Brazil, RNP is investing in the deployment of a 6,200 km optical route from Fortaleza to Porto Alegre,34 passing through many state capitals. (Figure 9.11). To carry out this project, RNP has the support of the Ministry of Science, Technology, Innovations and Communications (MCTIC). Other partners of BELLA in South America seek similar support in their own countries.

FIGURE 9.11. Likely trajectory of the Fortaleza – Porto Alegre optical route

Source: RNP

34 The Porto Alegre – Buenos Aires route already uses scalable optical infrastructure, the result of an agreement between RNP, RedClara and Level3, which guaranteed the right of use for 15 years by RNP and RedClara of about half of the spectrum of the optical fiber along this route. In February 2016, a DWDM system with 10 Gbps technology was working along that fiber pair.


237

The route between Fortaleza and Porto Alegre is being built through swap agreements of RNP’s Redecomep fiber, exchanged for long-distance fiber from providers and telecommunications operators that have available fiber in parts of this route, as well as through partnership agreements with fiber owners for jointly lighting and sharing of the optical spectrum. Figure 9.11 illustrates the parts likely to be covered by this route. Lighting-up of this route will be carried out with 100 Gbps DWDM, providing an initial configuration of at least two 100 Gbps channels for RNP, one to meet RNP’s own require‑ ments for backbone capacity and one for BELLA. As this project is based on a scalable infrastructure, additional 100 Gbps channels can be provisioned as needed (Figure 9.12). Since this involves either dark fiber or sharing spec‑ trum, in the future higher-order technologies may be used, for example, 400 Gbps per optical channel.

FIGURE 9.12. Future RNP backbone in 2017

Source: RNP

238


The 100 Gbps Southeast ring Deployment of the Fortaleza – Porto Alegre route will begin in the Southeast Region, with the creation of the following important routes: Rio de Janeiro – Belo Horizonte and Rio de Janeiro – São Paulo, with extension to Brasília, form‑ ing the Brasília – São Paulo – Rio de Janeiro – Belo Horizonte – Brasília ring. For RNP, a scalable fiber ring in the Southeast has been for several years both sought and needed, and it should be built in the first phase of the Fortaleza – Porto Alegre route. The ring should already be available in 2017, built through swaps of fiber in some stretches, together with the lighting-up and sharing of optical spectrum of fiber provided by one or more partners. Figure 9.12 illustrates a possible configuration of the backbone of RNP in 2017, with the deployment of the Southeastern optical ring, extending to Brasília, and the Fortaleza – Porto Alegre route, at 100 Gbps.

CONCLUSIONS RNP has been operating since 1992 with funds from the federal gov‑ ernment, to provide network services to the academic community in Brazil through its own (and sometimes shared) communications infrastructure, which functions at urban, intra-state, national (interstate) and international levels. This chapter has described some of the key factors that allowed RNP to carry out a continuous and significant increase in capacity and in the geo‑ graphical extent of its communication infrastructure, since its inception as a project in 1989. These factors include the verticalization of this infrastructure, focusing mainly on the acquisition of long-term optical assets (often fraction‑ al), and the search for partnerships with infrastructure owners or telecommu‑ nications operators. The verticalization of infrastructure has been important for achieving im‑ portant economies of scale, supplanting the original practice of leasing tele‑ communications services to the end user. In its place, RNP has internalized the management of its telecommunications infrastructure, so it can now itself provide the telecommunications services it used to buy. Overall, this means managing the use of optical fiber assets, which makes RNP a different cate‑ gory of player, no longer a mere consumer of telecommunications services provided by others. The responsibilities are greater, but the benefits come in the form of reduced costs and greater flexibility in planning future develop‑ ment. It is because of this change that partnerships have become important,


239

whether with other operators of the academic network or with owners of tele‑ communications infrastructure. Due to the nature of telecommunications, economies of scale are also listed among the benefits of partnerships, even with entities whose clientele is completely different from RNP’s. Dealing with technologies of telecommunications and of networks also involves renewal. Optical fiber infrastructure can last 25 years or more. But the equipment used to exploit this fiber, as well as for data communications, needs to be changed at shorter intervals. Long-term projects need to take into consideration the renewal of equipment and tools. These investments and partnerships seek to increase network capacity and to reduce the cost of services, bearing in mind the continuous increase in data flow through RNP’s infrastructure, with new users and increasingly demanding academic applications.

REFERENCES BARROS, M. R. X., et al. Soluções para redes ópticas avançadas: ProjetoGiga (Solutions for advanced optical networks: Project Giga). Cadernos CPqD Tecnologia, Campinas, v. 1, n. 1, p. 3760, jan./dez.2005. Available on ; Accessed: Jan. 19th, 2016. CARVALHO, M. S. R. M., A Trajetória da Internet no Brasil (The trajectory of the Internet in Brazil). Dissertation (masters) – UFRJ, Rio de Janeiro, 2006. Available on:. Accessed: Jan. 16th, 2016. DE LAAT, C., et al. The rationale of the current optical networking initiatives. Future Generation Computer Systems, n. 19, p. 999-1008, 2003. SCARABUCCI, R.R., et al. Project Giga – High-speed Experimental Network. In: First International Conference on Testbeds and Research Infrastructures for the Development of Networks and Communities. Trento: Tridentcom, 2005. Available at: . Accessed: January 19th, 2016. STANTON, M.A. Non-Commercial Networking in Brazil. In: Proceedings INET’93. San Francisco: INET, 1993. Available on: . Accessed: January 20th, 2016.

240


______. RNPng: a próxima geração da RNP. Available on: . Accessed: January 19th, 2016. STANTON, M.A., et al. Building Optical Networks for the Higher Education and Research Community in Brazil. COMNETS 2005 Boston. Available on: . Accessed: January 19th, 2016. ______. RNP: a brief look at the Brazilian NREN. In: Terena Networking Conference (TNC2010) Living the network life (Selected papers). Vilnius, Lithuania: TNC2010 2010. ISBN 9789077559208. Available on: . Accessed: January 22nd, 2016.

| 10 |

THE ROLE OF NIC.BR IN THE EXPANSION OF BROADBAND INFRASTRUCTURE AND IN SERVICE QUALITY IMPROVEMENT Demi Getschko [ Brazilian Network Information Center – NIC.br ]

O

INTRODUCTION

n December 29th, 2015, the set of points of traffic exchange (IXPs) op‑ erated by the Brazilian Network Information Center (NIC.br) reached a peak of 1.4 Tb/s (terabits per second). In June 2015, Fortune magazine (NUSCA; VANIAN et al., 2015) reported that NIC.br was the largest Internet exchange point in Latin America and one of the largest in the world (Figure 10.1). In fact, more than 1 terabit per second is globally recognized as a large volume of traffic and the evolution shows that NIC.br is growing at a good pace (Figure 10.2).

FIGURE 10.1. Daily traffic on IX.br, last week of December 2015 (terabits per second)

Source: IX.br.

242

demi getschko

FIGURE 10.2. Annual traffic 2009-2015 in IX.br

Source: IX.br

Before discussing the IXPs (PTTs in Portuguese), a necessary comment: the Portuguese acronym, PTT, can be misleading in international terms, since PTT can also mean “Public Telephony and Telegraphy” or “Push To Talk”. Thus, NIC.br chose to add another acronym to IXP, which was internationally unambiguous: “IX” (Internet eXchange). But: • What is the concept behind the IXPs and what are they for? • What other roles do the IXPs play? • Why did NIC.br get involved in establishing IXPs in Brazil? • What was the investment of NIC.br in establishing and operating these points of traffic exchange in the Brazilian Internet? • How did NIC.br begin to care about establishing IXPs in Brazil? • How could NIC.br invest in establishing and operating these points of traffic exchange in the Brazilian Internet? • What other activities did the IXPs carry out in 2015? • And what should be the next steps to accommodate the exponen‑ tial growth of traffic from the digital inclusion of the second half of Brazil’s population and the increasing use of video content? In this chapter we will try to answer these questions.

SOME HISTORY To understand why IXPs were set up, we need to go a little back in time. In 1988, the first Brazilian links to international academic networks were the National Laboratory of Scientific Computing (Laboratório Nacional de

the role of nic.br

243

Computação Científica – LNCC) in Rio de Janeiro and the São Paulo Research Foundation (Fundação de Amparo à Pesquisa do Estado de São Paulo – Fapesp), both linking the national academic network to the Because It’s Time Network (Bitnet), which allowed the exchange of e-mail and files. In a short time, many Brazilian academic institutions sought to connect to LNCC or Fapesp to be able to connect their researchers to colleagues and col‑ laborators from abroad. The network mesh spread, and other universities and laboratories throughout Brazil formed the Brazilian branch of Bitnet, which worked until the mid-1990s (KNIGHT, 2014, p. 26). Figure 10.3 shows the network in December 1991. It was clear that this growth would not generate a scalable structure, capable of expansion and appropriate management. Therefore, other structures gradu‑ ally emerged. For example, in 1989, in São Paulo, the three state universities, plus the Institute of Technological Research of the State of São Paulo (Instituto de Pesquisas Tecnológicas – IPT) and Fapesp formed the Academic Network at São Paulo (ANSP). Also in 1989, the National Research Network now called the National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP) was established and its implementation begun. It was clear that planning for academic networks with a dedicated backbone was a pressing need.

FIGURE 10.3. Bitnet network in Brazil in December 1991

Source: Stanton (1993)

244

demi getschko

To plan a robust and stable expansion, a network topology that supported traffic and joined the local sections of the academic network had to be con‑ ceived. So a structure with sufficient capacity to implement the desired topolo‑ gy had to be designed and links leased from telecommunications operators to set up a network backbone. At the junction points, routers would manage the traffic. Within this structure, this backbone, subnets and participating com‑ puters would then connect. In 1991, the RNP backbone began to be leased from Embratel, then a com‑ pany of the Telebras group. Note that the protocol used to transport content on these channels was determined by the network in question: in Bitnet the Remote Spooling Communications Subsystem protocol (RSCS) was used, but if it were a segment of the Internet, for example, the content would be trans‑ ported by TCP/IP. Operators provided the “clean channels” and the requested transmission capacity. The tenant network was in charge of defining how they would be “filled”. In 1992, the backbone designed for RNP began operating via Embratel leased connections (Figure 10.4).1

1

The development of RNP, from its early days until 2015, is treated in detail in Chapter 9 of this book.

the role of nic.br

245

FIGURE 10.4. RNP’s backbone in 1992

Source: RNP

When TCP/IP was established as the standard protocol abroad, telecom operators understood that a new form of business and services was emerging: instead of renting “clean” channels, operators would sell connections with TCP/IP already included; in other words, to provide access to the requested capacity to their own new TCP/IP backbones. The interconnection between the operators’ backbones would provide access to the Internet as a whole. Thus, customers no longer needed to “see” or even worry about the topology of the Internet. They just had to lease the necessary capacity and leave the topology and routing up to the operator. In big cities, however, there was a problem. Part of the contracted band‑ width could be used only to “cross the street” to go to companies located liter‑ ally “right next to us”, but done by another carrier. That is, part of the leased capacity that could take us to a service in Mongolia was being used to link us

246

demi getschko

to a neighbor. And if the neighbor was using another operator, that “cross‑ ing the street” could theoretically mean a trip to Miami, where traffic would change operators and begin its way back. The creation of metropolitan inter‑ connection points could eliminate this inefficiency: local connections could be solved locally, with stakeholders exchanging traffic directly in a local ex‑ change point, a metropolitan IXP. Thus, the IXP behaves as a “square” for the exchange of local Internet packets at very low costs and with high efficiency. It eliminates the unnecessary “tourism” of packages around the world (reducing costs and excessive exposure to “prying”), improves transit time and allows fast and simple contact between providers and consumers. Not located within a local point of presence of a telecommunications operator, the IXP is neutral as regards its structure. Some activities that work in symbiosis with the IXPs and add robustness and scale to other NIC.br projects and services have been added over the years and will be described below. Some stand out: the addition of copies (mirrors) of Internet root servers, allowing an independent and more agile resolution of names; the presence of servers of the Traffic Measurement System (Sistema de Medição de Tráfego – Simet) and the Network Time Protocol (NTP.br), so users in neutral environments can measure various characteristics of the traf‑ fic, distribute the Brazilian Legal Time and further stimulate traffic to be ex‑ changed, already in the new version of the Internet Protocol (IPv6), among others functions.2 As for its central role, IXP allows the interconnection of subnetworks (Autonomous Systems) that want to exchange local Internet traffic and there‑ by increase the efficiency, resilience and stability of the network. The IXP attracts content providers, e-commerce, and entertainment in various forms, from text to “streaming” of music or video, and others. With an IXP, the provider can be located closer to its consumers, improv‑ ing the quality of service perceived by them. And this use is reflected today in the peak of IXP traffic, around 10:00 p.m., when most of us are at home using online entertainment services. The previous section partially answered the first question: why did NIC.br begin to concern itself with setting up IXPs in Brazil? Let us continue to ex‑ plain how funds for the setup and operation of these points of traffic exchange in the Brazilian Internet were obtained, what their current business model is 2

For the introduction of IPv6, see Gagliano (2010) and Hinden and Deering (2006).

the role of nic.br

247

and what other activities and functions IXPs perform today. To address this, it is necessary to remember how Internet governance in Brazil evolved.

The registro.br and CGI.br In 1988, with the first Brazilian connections to global academic net‑ works, it was necessary to think of unambiguous and standardized schemes to name the machines that made up our network. This practical need for an orderly and hierarchical form led us to the .br registry application in the Internet Assigned Numbers Authority (IANA). On April 18th, 1989, the .br Country Code Top Level Domain (ccTLD) was delegated to the group oper‑ ating the national academic networks, even before the start of the connec‑ tion and subsequent migration of our network to the Internet, which would begin in January 1991. In 1994, in another action that recognized the good operation of Brazilian networks, IANA assigned us allocation of IP address‑ es version 4, as well as the registry of autonomous systems (Autonomous System Number – ASN). The registry of names under .br and, from 1994 on, the local distribution of IPv4 segments was carried out by the registro.br group within the nation‑ al academic network. Everyone who worked managing .br administration al‑ ways sought full compliance with the protocols and the spirit of the Internet. In the case of registro.br operation, that meant compliance with the Request for Comments (RFC) 1591 of the Internet Engineering Task Force (IETF) that, among other principles, followed the basic rule of “the first applicant for an available domain name should have priority in registering it” (POSTEL, 1994). At the time, the registration of domain names was done without cost, both in Brazil (registro.br) and abroad (internic.net). The registro.br, as in virtually every country code registry back then, operated the Domain Name System (DNS) as volunteer work, dedicated to service and the good of the network. But even in this historical and somewhat romantic and heroic phase, one should not neglect the “common sense”: ok, the first comer “takes it”; but how to deal with those who come with “bad faith”? The world is more complex and humans are very “inventive”. While the Internet was still unknown (except in the academic community and in some sectors of civil society), some “ideas” arose on how to take advantage of the wide availability of good names on the web. With the domain registration being done for free, and while sectors of industry and commerce did not pay attention to the new reality, some might take advantage of the registry and subsequent “sale” of a good name. The

248

demi getschko

registrant could profit from selling it to anyone who was interested – and of course there is no clear harm in it – but registering names and companies under .br to then force them to “repurchase” them was beyond the limit of ethics and good practice. In 1995, the name landscape underwent a radical change: the American domain, with free registry on internic.net, started being managed by an American company called Network Solutions, incubated within the National Science Foundation (NSF) to make the process more self-sustaining. After all, companies were increasingly interested in the Internet. Thus, the domain names were no longer free. Fees were charged on registration and per year thereafter. Furthermore, this strengthened the view that the registries oper‑ ated by Network Solutions were now “generic”. That is, any interested party, from anywhere in the world could pay the requested fee, register under .com, .net or .org. The massive arrival of the commercial Internet, in addition to “old” academic Internet, was then clearly outlined. In Brazil, this awareness also developed at approximately the same time. Sensitized by the expansion of the Internet in Brazil towards the commer‑ cial activity, the Ministry of Science and Technology (Ministério de Ciência e Tecnologia – MCT) and the Ministry of Communications (Ministério das Comunicações – MC) established the Brazilian Internet Steering Committee (Comitê Gestor da Internet no Brasil – CGI.br) in 1995, giving structure and support to registro.br activities. In its constitution, CGI.br explicitly recog‑ nized a fundamental feature of the Internet: it should have a multistakeholder management, i.e. Internet governance should involve the various sectors of the Brazilian society. Therefore, in 1995, three years before the creation of the Internet Corporation for Assigned Names and Numbers (ICANN), CGI.br had a multistakeholder structure, with participation by representatives of civil society, the private sector, academia and government. None of these sectors had the majority. CGI.br created a support structure for a previously voluntary activity, pro‑ tecting the Brazilian registry, without forgetting good Internet standards. In addition to initiating studies to make the activity self-sustaining, CGI.br es‑ tablished measures to limit speculation in name registration. In 1995, only legal entities could register under .br, and to curb abuse in the name registry, CGI.br decided to limit registration to one domain name per legal entity. In 1997, following what Network Solutions had already done two years earlier, it was determined that the registration of names under .br would not be free of

the role of nic.br

249

charge: there should be a fee for registration under .br to make the operation self-sustaining. Note that, unlike what happened in the US with the “privat‑ ization” of .com, .net and .org registrations for Network Solutions, CGI.br was always faithful to the tradition of the network and thus the operation was kept in charge of non-profit institutions (GETSCHKO, 1997). By 2005, Fapesp’s corporate tax registration number was used to receive .br registration fees so as to segregate these resources. In late 2005, NIC.br took over the manage‑ ment of the financial process too. The limit of one registration per legal entity (Cadastro Nacional de Pessoa Jurídica – CNPJ) was not only draconian, the restriction would seem without much sense for a company that wants to profit from domain registration (for example, allowing the registration of different brands of the same compa‑ ny). But the restriction did curb speculative registration. Without violating the right of the “first comer”, yet another protective initiative was undertaken: CGI.br was the “first to arrive”, protecting the names that were considered critical, for example, internet.br, or those that were on the list of “famous brands” from the National Institute of Industrial Property (Instituto Nacional de Propriedade Industrial – INPI). Swear words were also removed from the available list, words the use of which under .br did not seem appropriate. Thus, following the registration rules contained in RFCs (and especially RFC 1951), some specific protections for were created for .br. Internationally, in 1998, ICANN took the role and functions of IANA and new generic domains were bound to appear soon. The controlled transition of the re-registration under .br, as we saw, was free until 1997. It was completed in 2000. Soon the funds from registro.br made the operations of basic services on the Internet in Brazil self-sustaining. The activity then started to show a surplus. It was time to create an institution with its own legal personality, to relieve Fapesp of administrative functions and the nuisance of legal disputes over these domain names, since, at the time, litigants ended up including Fapesp among the defendants. In 2002, a non-profit institution was created in order to take over the administrative and legal role of Fapesp: NIC.br, the executive arm of CGI.br. But let us return to the year 2000. The mission to suggest good norms for the Internet was always the focus of the action of CGI.br. With the surplus, some activities could be consolidated. For example, connections were piling up in Fapesp’s office building and it was clear that a new structure should be created for bringing together in one place all the links of the state universities

250

demi getschko

in ANSP and all other institutions seeking access to the Internet and ex‑ change of local traffic. Non-governmental organizations, research institutes and even companies that brought a telecommunications channel to Fapesp, at their own expenses, could now exchange traffic there. Of this accumulation of connections the first implementation of a point of traffic exchange emerged, the IXP, within Fapesp’s facilities. Part of the traffic that came there and was intended for other institutions with local links could be routed immediately, without burdening the international or interstate connection, especially the backbone of RNP and ANSP. To cite another example of voluntary service, in line with best practic‑ es, a group within NIC.br Security Office (NBSO) enthusiasts helped take care of the network security. This was the nucleus of the future Computer Emergency Response Team (CERT.br). Another group of volunteers, more numerous and more active today, discussed technical standards of operation and network management: the Working Group of Engineering and Network Operations (Grupo de Trabalho de Engenharia e Operação de Redes – GTER). In 2000, the operation of central resources of the Internet in Brazil, like name registry under .br and distribution of IP numbers, demanded additional technicians and space no longer available in the Fapesp building. The solu‑ tion was to change the Data Processing Center (Centro de Processamento de Dados – CPD) and those related to registro.br to another address. A building with adequate facilities was identified and then a floor was rented to house the machinery and the technicians operating core .br structures. The registro.br team and the CERT.br team also moved there. In 2003, CGI.br was recreated by a decree of the Presidency of Brazil. With a new composition of 21 members and direct elections for represen‑ tatives of civil community, the new CGI.br could fully activate NIC.br and gain autonomy. In December 2005, a resolution gave NIC.br responsibility for registering domain names under .br and allocating IP addresses and au‑ tonomous system numbers.

NIC.BR ACTIVITIES – HTTP://NIC.BR/ Established in 2002 and operational in 2005, following guidance of CGI.br and the decree that reformatted this Committee, NIC.br started to work on dif‑ ferent fronts to return, in services to the Internet community, the funds collected from its registration function. In 1997, when the BRL was almost equal to the

the role of nic.br

251

USD, the registration fee accruing to .br was 50 BRL at the registration of a do‑ main and another 50 BRL of renewal fee per year (equivalent to what was charged by Network Solutions). By 2003, the cost of registration had declined significant‑ ly, to 30 BRL per year without charge upon registration. This same fee in BRL was still being charged in 2016, although it represents only about 10% of the real annual value of 1997. It was therefore thanks to the growth in registrations, not in the amount charged, that NIC.br could achieve a surplus over its costs and use this surplus to finance activities related to the Internet in Brazil. The decree creating CGI.br specifically states that it is up to the Committee to generate good standards for the Internet in Brazil, provide statistics that show network expansion in various sectors and ensure the quality and safety of the network, in compliance with internationally recognized standards. This association between a multistakeholder Committee – CGI.br – and an operational non-profit agency that generates revenue to revert to Internet development in Brazil – proved a paradigm of success and a unique interna‑ tional case. Private resources collected in the registration and distribution of IP numbers could be used in activities to improve the Internet. The CGI.br/ NIC.br example has been praised countless times in international forums as a model to be followed. We can briefly list some NIC.br activities to meet goals outlined by CGI.br and enabled by the funds collected:

Registro.br – http://registro.br/ Registro.br is the central unit of NIC.br, responsible for revenue genera‑ tion. It exists ever since 1989 and, besides taking care of registration, publi‑ cation and maintenance of domain names that use .br, it also manages and allocates IPv4 and IPv6 (Internet Protocol versions 4 and 6) and registers autonomous system numbers (ASN).3 The publication on the Internet – in an agile, safe and timely manner – of the .br zone (file containing all names ending in .br) is made by identical 3

An Autonomous System (AS) is a subnetwork that has its own independent administration and is internally connected. The subnetwork of an AS contains a collection of routing prefixes advertised internationally by a standard protocol, the BGP-4. An AS may be under the operational control of one or more network operators. It has its own policy of internal routing and a common, clearly defined policy of external Internet routing. Each AS has a unique number (ASN) that identifies it. A description of the Autonomous System is available on: . Accessed: January 12th, 2015.

252

demi getschko

sets of servers and network equipment, managed and controlled by registro. br, located in São Paulo and several other cities in Brazil, as well as in San Francisco, US; Frankfurt, Germany and Seoul, South Korea. Thus, .br has servers in every region of the world. The spirit of the Internet proves alive when, for example, the German NIC (Denic) allocates servers within NIC.br, in response to our allocation of .br servers there. NIC South Korea (Nida) does the same, putting their machines in São Paulo, always free of charge between partners. It is also worth noting that .br was the second country code domain in the world after Sweden to implement an IETF specification suite called Domain Name System Security Extensions (DNSSEC). DNSSEC enables secure res‑ olution of domain names, as it makes this translation with the inclusion of a specific digital signature. This guarantees the non-interference of anyone seeking to maliciously corrupt the translation process. More than 800,000 of the 3.7 million .br domains are assigned today, which is very auspicious. Registro.br maintains a service desk available 12 hours a day to answer questions, solve problems and assist users in relation to the registration of domains under .br. Registro.br also works as the back end of some recently introduced generic Brazilian registries, like .rio, .globo and .uol. NIC.br participated in allocating new generic records of ICANN through registro.br and managed two new endings for its own use: .bom and .final.

CERT.br (Computer Emergency Response Team – Centro de Estudos, Resposta e Tratamento de Incidentes de Segurança no Brasil) – http:// cert.br/ CERT.br is one of the oldest units of NIC.br. It dates back to 1997, shortly after the creation of CGI.br itself. In addition to frequently participating in world security forums and having its work internationally recognized, CERT. br acts as a central point for security incident reports in Brazil, providing co‑ ordination and support in the response process and undertaking activities to tackle risks and prevent spam. It also generates and maintains a number of publications, practices and suggests new security activities. • Internet security booklet and fascicles Produces documents, educational materials, websites and portals that provide recommendations and tips on Internet security (password

the role of nic.br

253

protection, the fight against spam and other topics) for users and net‑ work administrators. • Secure Internet Portal (Portal Internet Segura – http://Internetsegura.br/) Centralizes the major Internet security initiatives in Brazil, helping Internet users to find information of interest and encouraging the safe use of the Internet. • Security courses Provides regular security training courses, attended by people in charge of this topic in various institutions. The course is backed by its equivalent from the Carnegie Mellon University and grants a certifi‑ cate recognized by this American institution. • Activities to combat risks and prevent spam In this area, CERT.br monitors the origin/destination of the spam in the Brazilian part of the network. After studies, it recommended the control and blocking of port 25 (TCP-SMTP) for household Internet access. This initiative generated a CGI.br resolution supported by the Brazilian reg‑ ulator, the National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel) and the operators, resulting in a significant reduction of spam originating in Brazil. • Incident notification – http://www.cert.br/stats/ Centralizes a large collection of data categorized by type of incident and date, showing the most frequent and worrisome for each period.

Ceptro.br (Center for Studies and Research in Network Operations and Technologies / Centro de Estudos e Pesquisas em Tecnologias de Redes e Operações) – http://ceptro.br/ It handles projects mainly related to network technologies and Internet operations in Brazil, aiming at its development and continuity. Its main proj‑ ects are: • Metropolitan Traffic Exchange Points – http://IX.br/ Provide infrastructure for direct interconnection between networks of different Autonomous Systems, improving the quality of the Internet and adding robustness, agility, response time, resilience and geo‑ graphic organization, thereby reducing investments and operating

254

demi getschko

costs to the AS. Originally called PTT.br and today also known as IX.br, this program installs, maintains and promotes the necessary metropolitan infrastructure in 25 locations in the country for direct in‑ terconnection between the participating networks (AS) that make up the Brazilian Internet. Other places for IXP installation are constantly being studied. The model rationalizes cost: a participating network can deliver its traffic as close as possible to the destination, with better performance and quality perceived by users of the service. Thus, a PTT.br (IX.br) is an interconnection within a metropolitan area, with different points of network interconnection (PIXes) and under the centralized manage‑ ment of NIC.br. From the beginning, NIC.br tried to disseminate the culture of creat‑ ing Autonomous Systems and their participation in IXP. Figure 10.5 shows the growth of AS in Brazil, representing 65% of AS of Latin America and the Caribbean (LAC) region.

Figure 10.5. Number of ASNs in Brazil, 1999-2015 Cumulative amount of ASNs with IPv6 and IPv4 4000 3500

Quant. ASN Quant. ASNs with IPv4 Quant. ASNs with IPv6

2500 2000 1500 1000 500

20 15

20 14

20 13

20 12

20 11

20 10

20 09

20 08

20 07

20 06

20 05

20 04

20 03

20 02

20 01

20 00

0

19 99

quant. ASNs

3000

Source: NIC.br.

• IPv6 Portal – http://ipv6.br/ This portal collects all the useful information to the necessary and ap‑ propriate transition to the new version of the Internet Protocol, the

the role of nic.br

255

IPv6.4 The portal includes online courses open to the public, a collection of rules and standards, information on the current state of migration and ways to approach it, especially in view of the already depleted stock of version 4 addresses (IPv4). In addition to textbooks, the portal deals specifically with the novelties that the Internet of Things (IoT) brings. Ceptro.br is also responsible for on-site courses on IPv6, which are given regularly and free of charge in São Paulo and other cities. About 5,000 professionals and multipliers have gone through classroom courses of the Study Center. • Brazilian Legal Time (Hora Legal Brasileira) via Network Time Protocol (NTP) – http://ntp.br/ NIC.br distributes free Brazilian Legal Time via NTP in conjunction with the National Observatory. The service aims to keep the devices on local networks and the Internet on correct time, which is a fundamen‑ tal precondition for interactions and possible security operations to be executed with confidence. The portal provides documentation and guidelines for proper installation of NTP on user devices. • Simet (Sistema de Medição de Tráfego – Traffic Measurement System) – http://simet.nic.br/ The Simet system, developed entirely in Ceptro.br, carries out 100% independent measurements of network connections quality, mainly to provide information to users, service providers and people responsible for the Autonomous Systems. This information enables monitoring and improving Internet access in Brazil. The Simet conducts perfor‑ mance tests on networks with access to the Internet through free soft‑ ware, available for anyone to use and also a specially developed device (the Simet Box), which allows constant operation and measurement. Simet uses servers installed within Brazilian IXPs. All its infrastruc‑ ture and operations are independent, neutral and of the responsibility of NIC.br. The system also works in mobile devices (iPhone, iPod, iPad and mobile devices running the Android operating system). The

4

The definition of version 6 of the Internet Protocol (IPv6) can be found in RFC 2460, “Internet Protocol, Version 6 (IPv6)” S. Deering and R. Hinden, December 1998. Available on: . Accessed: April 1st, 2016.

256

demi getschko

http://simet.nic.br/mapas/ website offers a dynamic mapping of the results achieved throughout the country to date. • Zappiens.br – http://zappiens.br/ Zappiens.br is a platform for storage, dissemination and distribution of content in Portuguese, especially video. It is an initiative that shares its name, concept and intent with its namesake in Portugal – the Zappiens.pt (http://zappiens.pt/).

Cetic.br (ICT Study Center – Centro de Estudos em TIC) – http://cetic.br/ Cetic.br conducts research on access and use of the Internet in Brazil, facilitating the formulation of public policies. Based on its research, Cetic. br monitors and evaluates the spread and socioeconomic impact of ICT. Indicators for more than ten years concerning the Brazilian Internet are avail‑ able on its portal, including Internet penetration statistics for households and businesses, web access by children and adolescents, e-government, civil so‑ ciety and healthcare institutions. In 2015, a tool to consolidate and visualize interesting data was added to it. The results of Cetic.br are obtained in compliance with international standards and the norms of the Organization for Economic Cooperation and Development (OECD), the International Telecommunication Union (ITU) and UNESCO. The work of Cetic.br is recognized internationally, so much so that it received the UNESCO seal of “Regional Center of Studies for the Development of the Information Society”.

Ceweb.br (Research Center on Web Technologies / Centro de Estudos sobre Tecnologias Web) – http://ceweb.br/ Ceweb.br is the newest member of the set of services provided by NIC.br. It aims to disseminate and promote the use of open web technologies. Ceweb. br also seeks to promote and drive the evolution of the web in Brazil, develop‑ ing studies, research and trials of new technologies. Ceweb.br’s origin is related to NIC.br’s hosting of the regional office of the W3C (World Wide Web Consortium), an organization that works actively to create and disseminate standards for the web, including efforts to promote access for everyone.

the role of nic.br

257

Good Practices Portal (Portal de Boas Práticas) – http://bcp.nic.br/ Brings together recommendations produced by the technical team of NIC. br on good practices related to a particular technology or tool.

WHAT THE FUTURE HOLDS Registration under .br, with the proceeds from the distribution of IPv4 and IPv6 numbers, in addition to Autonomous System Numbers, enabled NIC.br to invest in activities to benefit the Internet in Brazil. The activities and programs listed above require investments and operational resources. In addition to these programs, NIC.br also allocated resources to promote four international meetings: two meetings of the Internet Governance Forum (IGF) in 2007 in Rio de Janeiro and in 2015 in João Pessoa, and two meetings of ICANN, in 2003 in Rio de Janeiro and in 2006 in São Paulo. The IXP is undoubtedly one of the most important investments, both in its traffic exchange role and in its role as “neutral point” for Simet measurements. The São Paulo IXP, which accounts for 80% of the total traffic, is particularly important because it brings together more than 700 Brazilian Autonomous Systems in 29 Interconnection Points (PIXes) scat‑ tered throughout the metropolitan area. The arrival of so many connec‑ tions in the NIC.br facility, each one represented by two separate optical fibers and, in general, via aerial routes (on poles), merits attention (Figure 10.6). After all, a simple car crash can affect a utility pole and thus prevent a dozen fibers from reaching their destination. Or an incident in NIC.br’s own building might cause some unavailability of the facilities there. The various central services of NIC.br, such as registration and publication of .br area, are highly redundant, with replication in the country, in the world and full availability over all these years. Even in the case of IXP, there are partial redundancies in place, but given the physical characteristic of this concentration of lines, buying a high-performance (and more expensive) router is not enough to achieve final redundancy.

258

demi getschko

FIGURE 10.6. Links between PIXes (São Paulo)

Source: IX.br. Available on: . Accessed: Oct. 8th, 2015.

Thus, the essential role played by IXPs to support the Brazilian Internet, the traffic growth in recent years and that expected in the future require constant review of the structure, quality and quantity of IXPs operated by NIC.br. Having ever more videos and entertainment online, the gradual migration of commer‑ cial usage spikes to a figure with almost permanent heavy traffic and the need to reduce delays and latencies all indicate the need for continuous efforts. The popularization of online entertainment, with increasing domestic consumption of videos, becomes clear when we notice that the peak usage stabilizes at 10:45 p.m. The trend is that consumption will continue to in‑ crease rapidly. Thus every effort to build a building to house an IXP like São Paulo’s, the peak usage of which is above 1T/s (terabit per second), is justified. Increasingly powerful equipment, fiber channels with higher capacity, op‑ tical modulation multiplexing using Dense Wavelength Division Multiplexing (DWDM) are needed, even if they are quite expensive. Therefore, providing for expansion with plenty of capacity to meet São Paulo’s IXP future demand required acquiring its own piece of land, on Av. João Dias, in the southern region of the city, in a high elevation area with ap‑ propriate dimensions. There a data center that safely hosts both the growing metropolitan IXP of São Paulo and the central NIC.br services was built. The data center also includes facilities for monitoring, operations and a service

the role of nic.br

259

desk. The building, financed entirely and directly by NIC.br funds, was inau‑ gurated in November 2015. It has more than 3,000 square meters of space, full redundancy of generators, storage tanks for 5,000 liters of diesel fuel, a green air conditioning system with triple redundancy and a dual optical fiber ring (also built exclusively by NIC.br) for its interconnection with the facilities in the original building on Av. Nações Unidas. The area that will be occupied in the Data Center on Av. João Dias is six times larger than the current facility. There will also be enough electricity and air conditioning for future expansion. Additionally, dedicated servers, partici‑ pating in the IXP providing content to end users, will also be allocated there. We can only hope that NIC.br’s income will grow at the same pace, al‑ lowing the implementation of the support plans. Even with the arrival of new generic domains, the trust of Brazilians – always demonstrated by the use of .br as their favorite “last name” – will enable the continuation of activities that support the evolution of the Brazilian Internet, allowing CGI.br to expand its activities in favor of an open, neutral, secure and inclusive web.

CONCLUSIONS AND RECOMMENDATIONS We have shown how the income from Registro.br has enabled CGI.br, via NIC.br, to perform activities for the benefit of the Internet in Brazil. This is certainly something that cannot be neglected, so we should always seek additional funding mechanisms, aiming to provide better and more enduring support. NIC.br continues to invest in equipment and facilities. But part of the funding may also have to come from the participants. Today this is already happening indirectly, since it is up to each participant to fund their transpor‑ tation to the IXP structure; after all, the optical cables connecting the PIXes to the IXPs are financed and provided by managers of PIXes. NIC.br is not charged for the allocation of their IXP supporting equipment. The next steps that need to be taken certainly include providing more communication mechanisms between those institutions participating in the IXP, more flexible ways to test and add new connections and an ongoing dis‑ cussion about how to support NIC.br activities in this area. Finally, three important features are needed to assure the continued healthy development of the Brazilian Internet. First is the support of state and federal governments, as well as agencies like Anatel, reinforcing the neutral and free character of the telecommunications system over which neutral and

260

demi getschko

free Internet traffic moves. Second is the recognition of the multistakeholder mechanisms for construction and operation of independent traffic exchange points. And third is the important role played by NIC.br, not only to assure network quality but also to provide measurement and evaluation mecha‑ nisms. We have a good example of laws such as the Brazilian Internet Bill of Rights (Marco Civil da Internet). From the beginning, we have had a system of reciprocal obligations. For example, NIC.br has deployed domains like jus. br (Judiciary), leg.br (Legislative), mp.br (the Office of the Public Prosecutor – Ministério Público). For them, NIC.br provides free domain registration and receives, in return, greater visibility of the .br “surname” for the whole com‑ munity. CGI.br, an internationally recognized multistakeholder structure, will maintain the necessary conditions to support the growth of the Brazilian Internet and protect its essence.

REFERENCES GAGLIANO, Roque. IPv6 Deployment in Internet Exchange Points (IXPs). RFC 5963 (Informational), 2010. Available on: . Accessed: February 13th, 2016. GETSCHKO, Demi. Regras para a Distribuição de Números IP no Brasil. São Paulo, 1997. Available on: . Accessed: Feb. 13th, 2016. HINDEN, Robert; DEERING, Steve. IP Version 6 Addressing Architecture. RFC 4291 (Draft Standard), 2006. Available on: . Accessed: February 13th, 2016. IX.BR. Traffic. Available in: . Accessed: Dec. 31st, 2015. KNIGHT, P. T. The Internet in Brazil: Origins, strategy, development, and governance. Bloomington, IN: Author House, 2014. NUSCA, A.; VANIAN, J. et al. These are the buildings that make up the ‘cloud’. Fortune 8 June 2015. Available on: . Accessed: February 13th, 2016. POSTEL, J. Domain Name System Structure and Delegation RFC 1591, 1994. Available on: . Accessed: February 13th, 2016. STANTON, M. Non-Commercial Networking in Brazil. In: Proceedings of INET 1993, p. GFB1GFB.

| 11 |

THE CONTRIBUTION OF SMALL AND MEDIUM OPERATORS AND THE ROLE OF ABRINT Basilio R. Perez Breno Vale [ Brazilian Association of Internet and Telecommunications Providers – Abrint ]

B

INTRODUCTION

razil has a very special characteristic that only exists in a few countries in the world: the existence of thousands of small and medium-sized Internet service providers (ISPs). This chapter describes the trajectory of small and medium-sized ISPs in Brazil and their contribution to the expansion of broadband Internet service in the country. We also review some of the history of the Internet in Brazil and how some legislation, almost by accident, promoted the creation and growth of this segment of operators. We also describe the emergence and role of the Brazilian Association of Internet and Telecommunications Providers (Associação Brasileira de Provedores de Internet e Telecomunicações – Abrint), an organization that brings together small and medium-sized operators. Finally, we offer our thoughts on the current market dynamics and public policy to support small and medium-sized telecommunications companies, which play an important role in the regional expansion of Internet access in Brazil. Telecommunications regulation in most countries has followed a format that favors the creation of large economic groups, which almost always results in domination of a few telecommunications companies that then form oligop‑ olies. This stems from the concept that in the telecommunications industry firms need to be large to be economically viable. This was true in telecom‑ munications formats prior to the emergence of the Internet. Thus, regulators created conditions for concentration of areas that could only be covered by large economic groups. The TCP/IP protocols used on the Internet, with their basic concept of routing data packets, provide numerous possibilities for the

262

basilio r. perez • breno vale

existence of small and very competitive businesses, occupying niche markets or serving areas of little interest to large economic groups. Even in countries with a more favorable competition regulation for small and medium-sized enterprises there is a prevalence of conditions permitting the emergence of a few powerful groups that dominate telecommunications, reducing direct competition and, consequently, the provision of services in areas of little economic interest. In this respect, Brazil – because of several historical factors to be dis‑ cussed below – allowed the existence of thousands of small and medium-sized telecom operators. The existence of these small businesses prevents market domination by big companies. This ended up automatically regulating prices thanks to the simple existence of direct competition. The other important con‑ tribution of small and medium-sized enterprises was to take Internet access to regions and places of little economic interest to large companies. This situation, however, is not due to a strategic planning by the govern‑ ment, but rather to natural market reactions to regulations and policies, as well as some actions taken by the government and the large economic groups. These small and medium-sized ISPs were the true pioneers of the Internet, the trailblazers of this market in Brazil. They only appeared because of subtle changes in regulation: specifically, the access providers in Brazil, which later became these small telecommunication companies, emerged thanks to a sim‑ ple act of the Ministry of Communications, which created the Internet Access Provider, with the specific name of Internet Connection Service Provider (Provedor de Serviço de Conexão à Internet – PSCI). The act established that this service would be classified as a Value-Added Service (Serviço de Valor Agregado – SVA) and, therefore, it would not be considered a telecommunications ser‑ vice. At first, this kept large telecom companies away from this market. Based on this act, telecommunications services were needed as a means for the pro‑ vision of other value-added services, including Internet access.

A SHORT HISTORY OF SMALL AND MEDIUM-SIZED INTERNET ACCESS PROVIDERS On June 1st, 1995 Rule 04/95 was issued by the Ministry of Communications. It clearly separated Internet services from telecommunica‑ tions services and allowed the emergence of thousands of ISPs. As mentioned above, the main merit of this rule was to functionally separate the Internet

THE CONTRIBUTION OF SMALL AND MEDIUM OPERATORS

263

from telecommunications, establishing that telecommunications are a medi‑ um on which to develop value-added services, including Internet access ser‑ vice. In a surprising vision of the future evolution of the market, this rule fore‑ saw the separation of physical access (telecommunications infrastructure) and the various players of the Internet environment in levels or layers of action. In this concept, valid until today, the Internet basically has three layers. The first is the physical transport media, almost always a telecommuni‑ cations service in the conventional format of data communication. In other words, telecommunications companies provide transport and physical con‑ nection between the various players, linking end users to access providers and these providers to other providers or other networks. The second layer of the Internet is made of the technical services that enable the proper functioning of the network; these services are known for some of their acronyms: DNS, FTP, NTP, SMTP, POP3, IMAP, HTTP etc. These are the various services of the TCP/IP protocols. In this second layer, the Internet access provider uses equipment to process service requests made by users of the equipment and perform the necessary operations according to the characteristics of each of the protocols of the services requested by the user. Thanks to these services, we can translate “what” and “how” users want to access the Internet, make such requests in the appropriate protocols, route those requests through the Internet to the final destination and recover the result of the response, handing it to the requesting user. The third layer of the Internet is the content, currently known as Over The Top services (OTT), which have existed since the early days of the Internet with the provision of access to web pages and repositories of files and documents. The ISPs, within the concept of value-added service, always acted in the second and third layers of the Internet. Before the publication of the 04/95 rule, which established the PSCI, in a country where telecommunications were a state monopoly and extremely inefficient, the arrival of the Internet would be bound to be maintained within that inefficiency scenario, with a monopoly that the federal state enterprise, Embratel was preparing to exercise. The government prepared to deal with the Internet through that state-owned long-distance telecommunications compa‑ ny. Access would happen through a system known as the National Packages Network (Rede Nacional de Pacotes – Renpac), making use of a dial-up access system centralized on servers in Rio de Janeiro.

264


The publication of rule 04/95, which is now considered as the birthday of the commercial Internet in Brazil, completely changed this scenario, prevent‑ ing Embratel from directly providing this new service and creating the figure of the ISP. Since Embratel and other telephone companies at that time – all of them state-owned – were focused on telecommunications services, value-add‑ ed services were to be provided by other companies. This enabled the emer‑ gence of the dial-up ISPs, mostly small private enterprises. Since telecommunications services (the first layer of the Internet) were already being supplied by other companies, access to the Internet could be provided over this existing infrastructure. For this reason, only a few servers and routers were necessary. They could be easily accommodated in a small data center with several telephone lines to receive the access of users and one or more links (connections) to the Internet for direct access to network resources. This enabled the emergence of many entrepreneurs who became dial-up access providers. The conditions for these pioneers were not easy, because at that time the cost of dedicated connections to the Internet, mostly provided by Embratel, was extremely high, despite the low speed of these connections. The overall speed of theInternet in Brazil (access to the international Internet backbone) in 1995 was only 6 Mbps. ISPs started their operations with only 64 Kbps or 128 Kbps links. Today most Internet home connections have more speed than the whole country in 1995, so it is difficult for those who did not live through that time to understand how it could even work. In fact, it only worked because of the format and the existing type of con‑ tent. At that time, the content was basically text with a few photos. There was no video or audio, because this would make access impossible at that time. The equipment was also quite expensive and not available on the Brazilian market, so everything had to be imported.

THE CONTRIBUTION OF THE FIRST PROVIDERS TO TRAINING AND GUIDING USERS Probably one of the biggest obstacles that those pioneers had to face in 1995 and 1996 was the creation of an Internet culture, since users had no idea what the Internet was and what could be done with it. Providers had to install access programs on clients’ computers, teach them how to use them and show some of the things they could do with that novelty.


265

A major concern for those first providers was that the Internet was just a fad and would end suddenly, so providers were concerned when customers did not access the web for a few days. So they acted proactively to encourage them to use the Internet. In this context, it is worth reflecting on the discus‑ sions currently taking place concerning the impact of OTT services on the use of the web. In the early days of the Internet, the concern was precisely that there were no major appealing services online. The web pages and the services available were almost exclusively used for school research or monitoring online news. There were no social networking applications, on-demand videos, movies or TV series as there are today. Now telecom operators complain that these OTT services demand a lot from their networks, requiring constant investments to accommodate the speeds necessary for these services. But in the past the networks had no ser‑ vices that created much traffic, because content was still very limited. That was why the providers were concerned, because if no significant content were created, the use of the Internet could be just a passing fad. Unlike what major carriers say today, these companies should be thankful for the existence of OTT services. Instead of simply wanting a piece of the revenues of these services, they should upgrade their networks to deal with the new realities and charge end users appropriate prices. Demand for the Internet is not generated by the existence of telecommunications networks, these networks are just the roads on which traffic will flow. Infrastructure alone is not enough, like roads. It is necessary, indeed indispensable to have something to carry on these roads and, in this context, the OTT are not the enemies, but great allies, because they create the demand that will need to be met by telecommunications carriers or operators, either large or small. Returning to the beginning of the Internet in Brazil, at that stage there were only small providers and a few medium-sized ones. Access was exclu‑ sively dial-up, and speeds were low. This is now known as narrowband, as opposed to the broadband service. As mentioned above, providers needed to teach users how to use each of the new developments that arose. One of the most important services of providers at that time was the technical support via telephone, depending on the level of the users. That is, these providers have helped to create a very important knowledge base with this constant training. Nowadays it is difficult to understand, especially for the younger gener‑ ations, the level of difficulty that existed between 1995 and 1996 to connect

266


a new user to the Internet. The operating systems of the time already had programs for the use of modems for dial-up, but they had no dialer program built into the system, let alone a browser or an application for e-mail. At that time, the most widely used operating system among the few people who had computer at home was Microsoft Windows 3.11. Though already a system in graphic format, it had no native dial-up features nor a web browser. For every new client that the provider included on the Internet, it was necessary to send one or more disks (this was the most common means of file-sharing at the time, almost nobody had a CD player) with the programs for the dialer, usually the Trumpet program and the Netscape browser. We often did not know how to install these programs, despite the contribution of small and medium-sized operators. These providers often had to go to the users premises to install the programs so that they could finally become “new customers”. Those were really heroic moments and a lot of hard work was done by many pioneers of that time.

TECHNOLOGICAL ADVANCEMENT AND HOW THAT HELPED SMALL PROVIDERS Fortunately, the first major technology support came with the spread of the Microsoft operating system, Windows 95. For the first time an operating system already had a built-in dialer and a browser, although precarious and less efficient than Netscape. At any rate, it facilitated the activation of new users. With these innovations, technicians rarely had to travel to the users’ homes to install Internet access programs. At that moment, Internet usage boomed in Brazil, with the emergence of national content contributing to it too, with “Cadê?” (Brazilian attempt at a search engine, very popular at the time), newspapers and magazines, and the appearance of portals like UOL, BOL and the like. With this very rapid and unexpected growth of the Internet, in mid-1997 the major operators finally realized they were going to be left out of this market, so they began to move. Dial-up connections were no longer made in analog lines and modems, the maximum speed of which was 32 Kbps, so they began to be replaced with dig‑ ital modems with speeds of 56 Kbps. These changes increased the reliability of connections and Internet content started to become more complex and interesting.


267

Thus the first attempts at broadband Internet began to emerge, initially with some experiments with the Integrated Services Digital Network (ISDN) technology, which doubled the speed of the dial-up service.

THE EXPANSION OF THE MARKET AND THE EMERGENCE OF MAJOR OPERATORS But what really took hold in Brazil was the ADSL protocol, which allowed not only sharing the same physical medium of the metallic pair used for tele‑ phony but also the transmission of data without the line staying busy, as was the case with dial-up. With this technology, broadband Internet finally came to Brazil, although at first with speeds of 128, 256 or 512 Kbps, but at that time it was a huge difference. Unlike what happened in the United States and other countries where cable TV operators prevailed as broadband providers from the start, in Brazil that did not happen for various reasons. The main reason was regulation. Cable TV operators in Brazil were still restricted by law to a few cities and they were not initially authorized to provide Internet services. For this reason, their networks were unidirectional and not prepared for Internet provision, needing several technical changes to allow access. However, legacy telephone networks of the former Telebras system and new networks were expanded by concessionaires after privatization auctions were already able to operate with ADSL standards. All they needed was some investment in equipment and modems for end customers. Because of regulatory issues, telephony concessionaires could not provide direct access to the Internet and therefore used the concept of authentication, to be done by ISPs, which already had the expertise to do so, since this was already done for the dial-up service. This authentication service was done via the Remote Authentication Dial In User Service protocol (Radius), one of sev‑ eral services in the second layer of the Internet normally performed by ISPs. Therefore, despite the development of the infrastructure of telecommunica‑ tion networks, small and medium providers continued to play a key role in access to the Internet. It is important to highlight that authentication is just one of the services of that second layer of the Internet and not the main activity carried out by ISPs, like some lawmakers and regulators claim. Within the technical standards and Internet protocols, the Radius authentication is only the User Datagram

268


Protocol (UDP) service done on port 1812. In other words, it is just one of al‑ most 50,000 listed and available services in the TCP/IP protocols performed in this second layer of the Internet by access providers. Broadband Internet service provision in Brazil was, therefore, “shared” between telephony concessionaires and access providers in its initial stage of development. But there was a very strong feeling of distrust of providers concerning the monopolistic intentions of the major companies, especially because the business model of utilities did not really provide for distributing responsibilities and revenues. It was more of a “temporary adjustment” to the regulation at the time. The General Telecommunications Law (Lei Geral das Telecomunicações – LGT) expressly forbade concessionaires to offer any service other than the ob‑ ject of the concession. The Internet was a value-added service that needed to be carried out by a different company. So they created the user authentication element by access providers. But unlike other countries, such as the UK and other European coun‑ tries, there was no real unbundling of the networks. This unbundling would have allowed ISPs to provide user access to the Internet via networks leased from the telecommunications operators in the form of partnerships. This concept of unbundling of networks was only effectively implemented in Brazil in September 2013. That was too late, since the market had evolved and many networks were already being duplicated because there had been no unbundling step.

THE REACTIONS OF SMALL OPERATORS TO MARKET MOVEMENTS Since the beginning of the Internet in Brazil providers always kept in touch, either through listservs for technical discussions or attempts at associ‑ ation. The exchange of information between them was very intense and these concerns regarding the intentions of the large concessionaires were both re‑ current and the focus of much discussion and analysis. In this context, a concept repeatedly advanced by some providers, grew stronger: “Be the owner of your networks.” Providers realized that the last mile of access to the end user should be owned by them or they would be forced out of the business within a few years. Thus, in the late 1990s, they began the first experiments with wireless access, initially using computers with adapter Wi-Fi cards and some routers


269

manufactured for indoor use and adapted for external use, in a typically Brazilian quick fix solution. Dial-up providers that began in 1995 were at that time becoming small telecom operators for their users also in the last mile. The regulations at that time still did not provide for this type of service. For Anatel and for the government in general, the only existing telecommu‑ nications companies that mattered were the concessionaires and a few “little mirrors” (espelhinhos) for telephony, (e.g. Intelig, Megatel, Vésper) that had been authorized in auctions conducted by the government.1 But in August 2001, Anatel created the regulation for Multimedia Communication Service (Serviço de Comunicação Multimidia – SCM) that allowed the regularization of the service that was being performed in an amateurish fashion by ISPs. This regulation was not, however, conceived to support the ISPs. It was designed to establish conditions for the creation of telecommunications sub‑ sidiaries of major carriers and concessionaires. They would be completely free to provide access to the Internet without the restraints of the concession con‑ tracts and the LGT. But the ISPs seized this opportunity and began to obtain their own SCM licenses, becoming telecommunications operators as well as ISPs. Thus, in addition to operating in the second and third Internet layers (content and services), the ISPs became telecommunications companies and began to operate in the first layer, building the last-mile physical access to the end user. In the beginning, there were only a few licensed companies, primarily due to interpretations of the regulation, which led to the understanding that an ISP with SCM license could provide communication services through part‑ nerships with other providers. Although this concept is really valid and con‑ sistent with the SCM regulation, some distortions ended up leading to what became known in the market as “rented license”. Anatel properly countered this distortion, in which some companies “rented” licenses to other providers. When it became clear to providers that the best option was for each company to have its own SCM license, the number of licenses began to explode. Only one of these “mirror” companies was successful: GVT, which was recently merged into Vivo (Telefônica group).

1

Espelhinhos was the generic name that the market created to designate authorized telephone companies created to ensure competition with companies that acquired the telephony concession through public auctions. The intention was to prevent a private monopoly from arising to replace the state-owned model.

270


By 2016 there were over 5,000 companies authorized to provide the mul‑ timedia communication service. Given the size of Brazil, the existence of this large number of regional operators is key to the growth of the network infra‑ structure and digital inclusion, especially in more remote locations that are not economically attractive for large operators. This period of strong expansion of companies with SCM licenses replac‑ ing the previous model of license sharing was very troubled and difficult. But telecommunications regulations have always been made thinking exclusively of the large operators in the market. To Anatel and the government in general, regional providers were irrele‑ vant and mostly irregular, especially in view of Anatel as a watchdog that fre‑ quently shut down providers. The small and medium providers have always had some irregularities, especially considering the intricate tangle of regu‑ lations designed for large companies, which had several specialized depart‑ ments just to comply with these regulations. Anatel was caught by surprise with the sheer number of regional providers that were obtaining their SCM licenses. There were major conceptual differences as to what should be the working models of these companies.

COMPETITIVE EDGE OBTAINED WITH THE EMERGENCE OF ABRINT In this context, in 2009 Abrint was founded. In its first event, which gath‑ ered more than 250 entrepreneurs in the sector, Abrint managed to have an Anatel representative as a speaker. It was of course a lecture full of concep‑ tual controversies. This made it clear how hard Abrint would have to work to consolidate the understanding of business models so that small and medi‑ um-sized enterprises could operate with legal security, within a strong regu‑ latory environment. One of the premises for the creation of Abrint was the need to establish dialogues with Anatel and all government agencies in an open and productive manner, especially to convey the message of how regional providers operate and their relevance and importance to the country. In the seven years that have passed since Abrint was founded, a lot has changed. The Ministry of Communications now understands the role of regional Internet providers and has offered strong support to the growth of this market segment, adopting in‑ centive policies, conducting frequency auctions in a municipalities format (so


271

that they are economically viable for regional providers), as well as financing projects and guarantee funds for loans obtained by the providers. Anatel also radically changed its view regarding regional providers. They are no longer seen as irregular companies and have begun to be seen as partners in Brazil’s development and the digital inclusion of its population. Thanks to this change of view, there were some regulations with the introduc‑ tion of concepts and definitions of small and micro telecommunication pro‑ viders. Under these new regulations, companies with up to 50,000 users are considered small providers of telecommunications services. If they have up to 5,000 users, they are micro providers. These definitions embraced all the regional providers, so Anatel began to issue regulations, taking into account these differences. Its requirements became more flexible and adapted to the size and capacity of these companies. Although at first sight it may seem that users might be negatively affected by this greater flexibility in the companies’ obligations based on their size, this does not actually happen. That’s because small businesses, to assure their own survival, have always been very close to their customers and offer advan‑ tages and forms of assistance that are almost impossible for a large corpo‑ ration to provide. For example, it is easy to see that it would be pointless to require these companies to have a 24x7 toll-free help desk, since in most cases customers personally know the owners of the providers and often have their personal phone numbers. In addition to the changes obtained with more favorable rules, both through a close dialogue with Anatel and the constant monitoring of the mar‑ ket and the environment itself created by the Internet, Abrint also brought about a very important change in the relations among regional providers. They no longer see themselves only as competitors and have recognized that they have many things in common, so they can share information to strength‑ en the whole group. Annual Abrint events have always sought to share a lot of information with regional Internet providers, not only in the technical area, but mainly to learn about market options and new business formats. Case studies of other regional providers presented during these events, with their successes and even some failures, have always enhanced the development of these compa‑ nies and strengthened the Internet market.

272


EXPANSION OF SMALL AND MEDIUM-SIZED OPERATORS In recent years, there has been a consolidation and an evolution of these regional Internet providers. Although there are still many companies that offer the service only via radio, over half of them already work with optical fiber and speeds usually found in more developed regions. There are some cities with less than 30,000 inhabitants fully wired with optical fiber cables deployed by regional providers, using their own resources and without direct support from public or private financial institutions. The most interesting thing about these operations in small towns is that they would be considered economically unattractive by major carriers, but they are a promising business to regional operators. This is precisely the main advantage of Brazil having these thousands of regional providers, because they create the possibility of digital inclusion precisely where there is no inter‑ est from large operators. This Brazilian reality also greatly facilitates the reduction of prices and largely avoids the creation of monopolies where the domination of major op‑ erators would cause an inevitable rise in prices and a consequent reduction in supply in places of little interest. When major carriers reach these places, currently served by regional operators, they will have to provide comparable prices to try to get a share of this market. This competition only benefits the end user.

CHALLENGES TO THE GROWTH OF SMALL AND MEDIUM-SIZED NETWORK OPERATORS In 2015, the market situation in Brazil was very worrying. In the case of telecommunications and the Internet, we did not notice any decrease in the demand, but the very instability of the economy affects the supply of credit for expanding networks and producing terminals for users, since tax reduc‑ tions for smartphones and notebooks were being removed. Regional provid‑ ers need clear definitions regarding the promise of a guarantee fund for loans with clear rules, appropriate to the size and reality of these companies. With the existence of this guarantee fund and a possible improvement in the eco‑ nomic situation in Brazil, which would allow the release of long-term loans under appropriate conditions, providers will be able to prepare business mod‑ els to install optical fiber in cities of less than 100,000 inhabitants scattered throughout Brazil. These are ambitious but perfectly achievable goals. Today


273

this is already being done on a smaller scale and with limited resources, and we are achieving great success. Another important possibility of growth can occur through the auction of municipal frequencies held by Anatel in the end of 2015. The result of the auction allowed the purchase by regional providers of frequency licenses for 2,904 Brazilian municipalities, over 50% of all mu‑ nicipalities in the country. In total, 314 regional providers participated in this auction, which actually offered the conditions and prices that we expected. It will therefore be an important lever for providers in cities where topology or population density hinder cabling through optical fiber. Regional providers will finally have frequencies free from interference to provide services with the quality needed to meet the demands of speed and performance of today’s existing applications and content on the Internet.

RISKS RELATED TO INTERNET GOVERNANCE As previously mentioned, the Internet consists of three layers. The first, the telecommunications layer, should be completely neutral by nature and definition. The connections between the various players should be restricted exclusively to the speed and quality of service levels that have been contracted by the users. The second layer of the Internet, technical services, is subject to great‑ er risks of interference by the regulatory agencies, because of its similarities to telecommunications. There are those who confuse these two layers, even intentionally, and try to create regulatory frameworks affecting services and protocols of the second layer. This is also because most of these services in the second layer are within a given geographical territory, which makes this interference in its governance by regulatory agencies and governments easier. The third Internet layer, the content or OTT applications, due to its glob‑ al nature, is extremely difficult to control and therefore does not facilitate a direct attempt at regulatory intervention. It is more subject to taxation than technical regulation. One of Abrint’s main goals is the defense of Internet governance in a mul‑ tistakeholder arrangement, like that of CGI.br, which has become a model for Internet governance worldwide. The risks posed by attempts to regulate the Internet services are very serious and began with attempts by the International Telecommunication Union (ITU) to gain power over the Internet in 2012.

274


These attempts have been dealt with, but not completely eliminated. They still need to be addressed because they could completely undermine the cur‑ rent operation of the Internet as we know it. It is vital that the Internet con‑ tinue to have a multistakeholder governance system, with no player (govern‑ ment, regulatory agency, industry, operators, academia, or any other) having full control over it. The most important thing is the diversity and joint gover‑ nance that exist today. The Brazilian Internet Bill of Rights (Marco Civil da Internet) that became law in April 2014 established the basic rules for the Internet in these three layers. In the first layer of the Internet – telecommunications – net neutrality has always been a key factor. There is not the slightest possibility of a telecommu‑ nication service working without neutrality concerning data packets that cir‑ culate within the telecommunications network, regardless of the technology employed. In the second layer of the Internet – service protocols – neutrality must also be respected. In this layer, neutrality must be guaranteed for each type of service, without discrimination of content or application. In this layer there may occur technical restrictions or prioritization of some services that use techniques known as QoS (Quality of Service). Some services are more susceptible to network latency and therefore need to be prioritized. For example, Voice over Internet Protocol (VoIP) applica‑ tions can be prioritized over e-mail, because verbal communication does not accept high latency and for e-mails this does not cause any damage. In the third layer – applications – it does not make sense to talk about neutrality, but the Marco Civil also established criteria and standards for reg‑ istration logs and privacy protection.

PROSPECTS FOR SMALL AND MEDIUM-SIZED OPERATORS With their optical networks growing by leaps and bounds within the mu‑ nicipalities, regional providers are organizing to deploy intercity and even in‑ terstate optical rings. There is a clear need to build networks independent from major carriers and allow several regional providers to connect, which will enable the exchange of traffic and decentralizing the delivery of audio and video content. These exchanges usually take place at traffic exchange points (IXPs) in state capitals. This movement, called “regional IXPs”, has been supported by the giant producers of content, such as Google, Netflix and


275

Facebook, since the closer to the user that content is, the better the connection and the quality of access to it. From the perspective of the providers, IP and transport links costs are reduced significantly, because much of the most used content is within the regional IXPs. The Ministry of Communications, in partnership with the National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP) and DynaVideo, a company based in João Pessoa, is participating in a pi‑ lot project to develop a national platform for content delivery – the Content Delivery Network (CDN) – to be used initially by the government. The idea of the federal government is to develop a public database that could bring this information from other sectors closer to the government or non-govern‑ mental organizations that need to access this public data. Several agencies have already expressed interest in using the platform, including Telebras, the Ministry of Science, Technology and Innovation (MCTI), the Ministry of Education (MEC), the Brazilian Communications Company (Empresa Brasil de Comunicação – EBC), and the RNP itself. The model, which can be imple‑ mented within the regional IXPs, will provide citizens with fast and quality access to government content and online education, for example. In the first decades of the Internet, the focus of regional operators certain‑ ly was not the long-distance transmission of content. Although this is a very important aspect for the network, it certainly will not be its only definer as we move toward a more connected future. But we need to see it as an increasingly present and inevitable movement. Therefore, we have to prepare the networks for large image transmission formats such as 4k or 8k. Another important issue in the coming years will be the inclusion of more people in the Internet. Much of the responsibility for this digital inclusion rests with regional providers. There are several movements for organizations such as the United Nations (UN) to establish Internet access as a fundamental right of human beings. Achieving this goal, that is, connecting everyone, will be a revolution of massive proportions. Currently, only 2 billion of the 7 billion people in the world connect to the Internet every day. The growth of this number has slowed, and today it is only 9% a year, according to the Broadband Commission of the United Nations.2 2 Available on: . Accessed: March 20th, 2016.

276


CONCLUSIONS The diversity of perspectives from all different parts of the globe, with peo‑ ple who interact and jointly deal with problems, can bring about even more changes than we expect today. Especially in the social field, the Internet en‑ ables discussions on regional and global issues, giving voice to all people. But in order to achieve that, access is paramount. Much of this diversity of experience is perceived and experienced in areas where network access is still precarious. This raises the importance of regional operators that take access to remote areas, poor neighborhoods, rural areas and towns with less than 50,000 inhabitants. Studying the Brazilian experience, remarkable for the presence of thousands of small and medium providers, may be useful for other countries with similar challenges.

| 12 |

THE CONTRIBUTION OF THE LARGE AND MEDIUM OPERATORS João Moura [Brazilian Association of Competitive Telecommunications Service Providers – TelComp]

I

INTRODUCTION

n this chapter we will review the development of telecommunications in Brazil, the events after the privatization cycle that began in 1998, the lay‑ ing of the foundations for development of broadband, and the present situation. We then take a look into the future. We will take as benchmarks the large and medium telecommunication providers, including the vertically and horizontally integrated operating companies and the operators specialized in market segments (specialized as regards products offered, the technology em‑ ployed, and their geographic presence and reach. We will then discuss the range of different types of enterprises that are members of Brazilian Association of Competitive Telecommunication Service Providers (Associação Brasileira das Prestadoras de Serviços de Telecomunicações Competitivas – TelComp), an entity founded in the year 2000 to encourage the development of competition in the telecommunications market after the privatization. Sectoral regulation plays an essential role in the development of telecom‑ munications. The regulatory framework defines or limits the areas and con‑ ditions for business development, not only by historic providers that already existed before the privatization but also by newcomers that were drawn in by the potencial of the market that opened its doors. Hence we will offer an overview of regulatory aspects that were and are still paramount for its devel‑ opment – initially in telephony – and evolution towards broadband in Brazil up to the present time. The chapter reviews the history of the sector’s regulation, highlighting the challenges faced by providers as they contemplate expanding their businesses through new investments in networks and services. The aim here is neither

278

joão moura

to undertake a legal analysis nor to delve into discussions on controversial as‑ pects of legislation and regulation that impact the sector. Rather we will focus on practical issues that impact investment decisions. We will also trace parallels between the law and regulation models im‑ plemented in Europe and in the US in an attempt to spot relevant issues for Brazil and compare the results reached in those markets that have influenced Brazilian strategy. To illustrate the analysis we will describe some operator categories, emphasizing the specialized ones that are less known to the gen‑ eral public but play a vital role in the development of broadband in Brazil. We classify as specialized operators those that operate in specific market seg‑ ments, be it in services or geographic regions, in contrast to the integrated ones with national reach. For instance, Orange Business (France Telecom Group) operates exclusively in the demanding corporate market. MLS also op‑ erates in this market and provides services to other operators, exclusively with wireless solutions currently only in Rio de Janeiro. Finally we seek to identify elements of public policies that could help Brazil develop quality broadband networks, a critical factor for enhancing the country’s competitiveness in the new global digital economy.

MARKET CONTEXT AND TELCOMP’S INSTITUTIONAL ACTIVITY

Origin and market context Telcomp was founded in January 2000, bringing together pioneer compa‑ nies that entered the telecommunications market right after the sector’s pri‑ vatization. TelComp is a non-profit entity that in 2015 represented 54 compa‑ nies (or economic groups) in the telecommunications sector, including fixed and mobile telephony providers, data communications, pay TV, VoIP, among others. The entity seeks to encourage competition in the telecommunications sector, contributing to the sector’s legal and regulatory framework. Below we discuss the sector’s regulatory environment and TelComp’s activity. The General Telecommunications Law (Lei Geral das Telecomunicações – LGT – Law n. 9472 of July, 1997) was the starting point for the sector’s shakeout in Brazil. From this watershed on, a predominantly competi‑ tion-driven private sector replaced the former state monopoly. The National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel) was created, charged with creating a new sectoral framework driven by free market competition principles.

THE CONTRIBUTION OF THE LARGE AND MEDIUM OPERATORS

279

Initially, the old Telebras was broken down and its subsidiaries were re‑ grouped in three subholdings that were shaped as regional operators (South and Center-West; Southeast and Northeast; and North and the state of São Paulo). Embratel was preserved with its concession for national and interna‑ tional long-distance telecommunications as were some small local operators. These four new companies were privatized. Then permits for the creation of four “mirror” companies were auctioned in the same areas of the original operating companies and the Cellphone Mobile Service (Serviço Móvel Celular – SMC) was also released, also in the format of concession. The concessions for mobile service were replaced by the current licences for Personal Mobile Service (Serviço Móvel Pessoal – SMP), with some changes in the scope of the agreements. The model for Commuted Fixed Telephony Service (Serviço Telefônico Fixo Comutada – STFC) with the operating groups of the original and the new authorized (mirror) was the starting point for the development of a competitive market. For the first time a client in the same region would have two alternative fixed and mobile service providers at his choice. At the moment of privatization and in the following years, when new per‑ mits were granted for service provision, there was huge unmet demand for basic fixed telephony services and the prices charged were quite high. For this reason, the permits for fixed service could be auctioned, generating revenue for the federal government. These factors attracted new investors that, nevertheless, depended on sec‑ toral regulation to launch and develop their new networks in a sector dom‑ inated by old monopolists. Regulation regardings access, interconnexions, dedicated leased lines, and numbering plans, among others, were matters that demanded specific rules for favoring service provision by the newcomers. Though foreseen in a clear, thorough and updated manner, the set of rules in the LGT demanded specific regulations that would be issued by Anatel. As the newcomers were not able to meet some of them, they faced insurmount‑ able obstacles to reach the market. It was clear that the old operators had little incentive to cooperate and Anatel proved to be sluggish and inefficient in removing bottlenecks and promoting competition. This was the context in which the new operators decided to create TelComp and so work together in order to promote development of the telecommunica‑ tions market based on free competition. In the beginning the first associates were the corporate network forerunners, then came the so-called “mirrors”

280

joão moura

and the “little mirrors” (small companies allowed to provide fixed telephony services in smaller regions) and the newcomers in the cellular mobile services, in the B band. The common interest, that marked the beginning of TelComp, was the regulation designed to open the market to new companies, following LGT’s guidelines that sought to increase the number of service providers and the opportunity for consumers to exercise their right of choice. Furthermore, at that time Anatel was in its initial stages, still elaborating its original regulations and dealing with two other matters: privatization and issuing SMC licenses.

Privatization Preparing and conducting the privatization transactions in Brazil were extremely complex tasks. The size of the companies, the insipience of the regulatory model, the uncertainties about the physical condition of the net‑ works and about the investments that would be required, among other fac‑ tors, and a delicate economic environment with a very limited capital market to boot, demanded an enormous effort to conduct the privatization process. Nevertheless, the privatization in Brazil failed to attract the major internation‑ al private operators and consequently control of the biggest telecommunica‑ tions companies ended up in the hands of operators of less significance on the international scene or of investors lacking any experience in the sector.

SCM license concession At the same time, cellular mobile service was advancing, initially within the old state-owned companies and later by means of new concessions to op‑ erate in the B band, which would turn out to be the direct competitors of the previously state-owned companies, that had already been privatized. The B band operators launched “green field” projects, starting with the concession for radio frequency usage. Everything else would have to be created while at the same time meeting coverage requirements foreseen in the concession agreements. Unlike the privatization of the old state-owned companies, the auctions of concessions for B band mobile services attracted some of the larg‑ est American and Canadian operators. The government at the time celebrated the billions of BRLs collected with the sales of “wind”, an expression coined by the Communications Minister at the time, Sérgio Motta, reflecting the suc‑ cess of the endeavor, at a time when the national accounts were very fragile


281

and the exceptional revenue generated by the auctions was of utmost impor‑ tance to the federal government. In this context, Anatel’s priorities seemed to have been: (1) preventing failure of the privatized public companies, puting at risk not only essential services (fixed telephony) but also the very concept of private agents provid‑ ing public services and (2) making sure that the new cellular mobile service networks would be implemented quickly and hence meet the existing, strong and repressed demand. These priorities, that monopolized Anatel’s attention, may have been the original reason for the agency’s limited willingness to promote competition in the Brazilian telecommunications market.

Paradigm changes On the threshold of the twenty first century, the world had already realized that the future of telecommunications would rely on development of new gener‑ ation networks, mainly optical fiber and mobile networks. The Internet and the TCP/IP protocols made possible the development of communication between computers on a much larger scale than its original academic applications, open‑ ing new horizons for business development. These factors drove entrepreneurs worldwide to invest in new networks, foreseeing spectacular opportunities. Progress in the physical world of networks did not occur in synchrony with the development of profitable business models, that is, the monetiza‑ tion of network assets did not take place in tandem with the cash flow of pioneering entrepreneurs, leading to the so-called “Internet bubble”. New telecommunications operators, equipment suppliers, banks and financial investors suffered huge losses and many of them, facing bankruptcy pro‑ ceedings, exited the market. Nevertheless, the real assets remaining after the “bubble” did not vanish. On the contrary, they were absorbed by other companies that only years later witnessed the consolidation of the data transmission market, beginning with the strengthening of application providers with innovative business models and a strong ability to generate profits and cash flow. The movement that took place on a world scale had parallels in Brazil, where many of the pioneering optical fiber networks were taken over by the incumbent operators and by new firms that entered the market, attracted by the growth po‑ tential of still unmet demand for conventional fixed telephony services.

282

joão moura

A brief background TelComp was constituted by bringing together telecommunications op‑ erators that held Anatel concessions and had as main objective promoting competition in a widespread and isonomic manner. This model is different from other associations that, include equipment suppliers, consultants, and lawyers as well as operators. The denomination “competitive” is a literal translation of the term used in the US and in Europe to characterize the new operators that emerged to compete with the old monopolies. In the US, after the sectoral reform in the 1980s, the Competitive Local Exchange Carriers (CLEC) appeared. They evolved and became providers of various telecommunications services with an emphasis on broadband for the corporate market. These companies formed associations such as The Competitive Communications Association (CompTel – www.comptel.org), inspiring the name TelComp. In Europe, the same association movement gave rise to the European Competitive Telecommunications Association (ECTA – www.ectaportal. com). In Europe, old state-owned operators started operating outside their countries of origin, now as competitive ones. British Telecom, Telefonica and France Telecom Orange are examples of old operating companies that have grown a lot as competitive ones in European countries. Telecommunications in Brazil already had institutional representation by means of the Brazilian Association of Telecommunications (Telebrasil – www.telebrasil.org.br) and its affiliated entities that have as associates mainly old state-owned operators, employers’ unions, equipment suppliers, and con‑ sultants, among others. The founders of TelComp understood the importance of creating an entity focused exclusively on the development of conditions for market expansion through the entry of new operators. A healthy competitive environment could only be created with a regulatory framework that reduced barriers to entry and enabled new investments. It was clear that the text of the LGT and Anatel’s regulations, though well conceived, were not effective in practice, not only due to the reluctance of the dominant companies but also to failures in supervision and the application of sanctions. TelComp is an independent entity without an employer union’s structure, and is hence funded exclusively by voluntary contributions by its associates, with equal voting rights and other governance mechanisms that prevent deci‑ sions being taken based on the economic power part of any associate.


283

TelComp’s mission is: “Promoting a fair and isonomic competitive envi‑ ronment in the provision of telecommunications services, contributing in a constructive and ethical manner to the sector’s regulation as well as promot‑ ing social responsibility and corporate citizenship.” The following lines of work “drive” its activities: • Defending Anatel’s competence, power and duty in order to imple‑ ment regulations to promote competition; • Acting to implement pro-competition regulations and supporting pre‑ ventive action to curb abusive practices; • Avoiding the concentration of networks resulting in exclusion of com‑ petitors in markets by the large operators that have significant eco‑ nomic power; • Fomenting investment and infrastructure implementation, ensuring a stable regulatory environment with consistent rules; and • Acting to reduce taxation of telecommunications services, especially broadband.

TelComp’s governance and organizational structure In TelComp each associate is entitled to a vote regardless of its size and there is no veto right. As a result the interests of the majority prevail and the op‑ erators’ economic power cannot change the association’s decisions. This mod‑ el is different from the one adopted by other associations in which economic size guarantees special rights in decision-making. So, even though all the large operators are members, TelComp can continue to act to defend a competitive envirnoment, always guided by the interests of the majority of its members and not the economic weight of certain associates. This governance structure is responsible for approving the association’s work plans and positioning. That in‑ cludes decisions on lawsuits and administrative actions in the spheres of Anatel, the Administrative Council for Economic Defense (Conselho Administrativo de Defesa Econônica – Cade) and other public administration bodies.

Major institutional achievements Over the years TelComp has been successful in many initiatives to real‑ ize its institutional goals, always using arguments that are technically solid

284

joão moura

and coherent with its principles. Some of these issues will be detailed in the following paragraphs.

Unconstitutionality of charging for land use rights Through many legal actions, including in the Federal Supreme Court (Supremo Tribunal Federal – STF), TelComp avoided municipalities charging fees for land use to implement networks.1

Number portability This means the possibility to keep the same phone number when switch‑ ing operators. This regulation, for which TelComp worked hard, allows new operators to overcome the barrier represented by the existence of a phone number which is already known by the client’s contact network, that discour‑ aged people from switching service providers. According to Anatel, as of early 2016, every three months about one mil‑ lion subscribers switch their operators of fixed or mobile telephony services, while keeping their same numbers.

Wholesale markets The regulation of wholesale markets, whereby operators may lease lines of third parties to constitute their own networks, was stipulated in the LGT, how‑ ever it was always a source of conflict between newcomers and the dominant operators. Leasing allows networks to be constituted more efficiently and to expand quickly, in theory contributing to a faster return on their investments.

Isonomy in the municipalities for infrastructure rollout In the beginning, competitive operators had to pay different prices for rights of way when deploying infrastructure. Today such differential pricing still hinders the rollout of new networks. Every year TelComp has represented its associates in many lawsuits to enable network deployments.

1

For example: the actions 005.2005.007/4836 RE 581.947 and 2002.001.0134673 AP 2003.001.12811 RESP 729.407RJ RE 494.163RJ


285

Mobile Virtual Network Operator (MVNO) TelComp promoted a debate on the feasibility of creating a virtual network operator model to promote service innovations. To that end, it organized de‑ bates and put forward proposals for regulations, even without support from traditional mobile operators. By means of the MVNO regulation, a company may be granted permit to lease the infrastructure of a traditional mobile oper‑ ator and provide services with its own brand.2

General Plan for Competition Goals (Plano Geral de Metas de Competição – PGMC) This plan was approved in 2012 after almost a decade taken up by bu‑ rocratic processes. The plan replaced administrative regulation, specific for each kind of service, with economic regulation, based on imposing asymmet‑ ric liabilities on operators with Significant Market Power (Poder de Mercado Significativo). This is critical for understanding the regulatory framework of the telecommunications sector in the country, as explained in Chapter 3.

Anatel/Aneel Joint Resolution3 to enable the usage of utility poles of power distribution companies by competitive operators The incumbent operators have always occupied space on poles paying very low rates based on old contracts, many of them from the time when not only the electric companies but also telecommunications operators were state-owned. With the arrival of competitive operators, power distribution companies became aware of the business potential and began charging rates that were often higher. The additional cost represened a competitive handicap for the new operators. Besides, each pole has limited space for usage by the telecom‑ munications companies and, in general, large operators occupy many attach‑ ment points, not leaving any room for newcomers. The Resolution tackles important issues, but it depends on technical solutions that allow cables of many companies on a single attachment point. The Resolution also stipu‑ lates that each economic group can occupy only one attachment point on a pole, making room for other companies. This issue is of crucial importance 2

From this MVNO regulation on, new virtual operators were created in Brazil, such as Porto Seguro Conecta (from insurance group Porto Seguro/Itaú and from Vodafone, both members of TelComp.)

3

Aneel/Anatel 004/2014 Joint Resolution

286

joão moura

since, without access to poles, it is impossible to build networks in a quick and cost-effective manner.

TELCOMP’S CURRENT PROFILE

Members and profiles of the competitive operators As mentioned above, in early 2016 TelComp had 54 operators or economic groups,4 four of which are described further below, including all large telecom‑ munications groups and many companies specialized in certain products or market segments, having different sources of capital – from individual busi‑ nessmen to global companies, including also companies that receive invest‑ ments from private equity funds. We can group operators that are members of TelComp into the following categories.

Large integrated groups They sell a variety of services on a national level: América Móvil Group (Claro, Embratel and NET), Oi, TIM, Telefonica Group (Vivo and GVT). This segment is characterized by complete offers in all segments of the residential, corporate and government markets. The critical for these groups is their scale and the capacity to offer service packages to retain clients, en‑ joying synergies in all aspects of the business: products, physical networks, corporate structure, marketing and commercial aspects, and bargaining pow‑ er. The Telefonica and Oi groups expanded networks and broadband provi‑ sion by acquirng many new generation operators in defensive moves. The América Móvil Group has grown with the expansion of cable TV networks, using Hybrid Fiber Coaxial (HFC) networks for joint broadband service pro‑ vision for the residential and small and medium business markets. Embratel seeks to expand focusing on the corporate market, adding capillarity to its long-distance and high-capacity networks.

Carrier’s carrier These firms provide network services to other operators: Ascenty, BR Fibra, Internexa, Netell and Vogel Telecom. 4

The complete list of TelComp’s associate companies is available on: . Accessed: Feb.11th, 2016


287

These are companies that initially provide “connectivity” for other tele‑ communications companies, leasing circuits to interconnect their multiple units or facilities, such as antennas for mobile service, ground stations, data centers, etc. They are also known as neutral network operators as initially they don’t vie for the final client. These companies’ challenge is to quickly expand their networks and gain scale. They face problems in getting access to poles, ducts, and trenches but also problems concerning rights of way to deploy net‑ works, especially areas along highways. The natural tendency of companies is to add services of higher added value, investing, to that end, in data centers, cloud computing, etc. In this segment are companies still controlled by their original entrepreneurs and others that received new capital from investment funds. In this category we also observe moves to expand by means of mergers and acquisitions, between new companies as well as between new and incumbent operators.

Corporate market, medium and large customers In this group we can mention: Americanet, Mundivox, Compugraf, IpCorp, Option, CMA, Tesa, 76 Telecom, Viacom Lig 16, Go! Telecom, Netserv Telbrax, Horizons and WCS. The corporate market for telecommunications services has always been attractive for new operators given that incumbent operators have not paid much attention to it. Thus there is an opportunity to provide differentiated services, offering better quality and/or lower prices. This category of providers, always numerous and in expansion, seeks to combine telecommunications with differentiated solutions for applications, such as unified communications, virtual PABX, cloud computing, dedicat‑ ed circuits, Internet access etc. First and foremost, these operators use their own networks but some of them lease lines from third parties and provide differentiated service, or better Service Level Agreements (SLAs) than those offered by the incumbent telecommunications companies for the same kinds of services.

Large corporate market or global companies Members of this group: BT Telecom, Level 3 and Orange Business. This segment is similar to the previous one. Its basic feature is the fo‑ cus on large companies with high-level technical requirements, especial‑ ly in terms of security. Besides advanced solutions, these companies offer

288

joão moura

globally-managed network services. So they provide companies with facilities in many parts of the world where they have services in the same standard at their disposal, often hired in a centralized fashion.

Corporate market with wireless solutions We can mention: Neovia, Unitelco, MLS Wireless. These are operators that use wireless technologies (radio) both in spectrum bands requiring Anatel licenses and in free usage bands. Wireless networks have quickly evolved through new generations of digital radio and Wi-Fi solu‑ tions. Ease of installation and service, technical features (low latency for exam‑ ple) and lower costs than the optic fiber networks create certain advantages for these operators. On the other hand, there are constraints – such as physical barriers between antennas (buildings for example) and difficulties in installing towers for antennas – that hamper large-scale growth of such services.

MVNO’s Porto Seguro Conecta and Vodafone fall in this category. These pioneering operators provide mobile communication service by leasing network and spectrum rights of third parties.

Fixed (residential and corporate market broadband) LTE 4G Here are included On Telecom and SKY. On Telecom and SKY are pioneers in deployment of the brand-new fourth-generation (4G) fixed wireless communications technology for both the residential and corporate markets.

Operators afiliated to the utilities operators Members of this group are NRT Noroeste Telecom, SAMM, CPFL Telecom, Cemig Telecom, Copel Telecom. The potential synergy between electric power distribution network in‑ frastructure and that of roads prompted investments, in a complementary fashion, in new telecommunications networks. Starting from the utilities’ in‑ frastructures it is possible to deploy telecommunications networks efficiently and at low cost. This model has had some success with companies that were sold for high values, such as the AES Eletropaulo network in São Paulo and Light in Rio de Janeiro.


289

COMPETITIVE OPERATORS AND INCREASED PROVISION OF TELECOMMUNICATIONS SERVICES The new entrants in Brazil, mostly associates of TelComp, have devel‑ oped new networks and services with quality and price differentials, aimed mainly at the corporate market. Companies in any economic sector, be it industry, commerce or services, now demand telecommunications services of high quality and reliability to increase the productivity and efficiency of their business processes and also to use new computer technologies, based on “cloud” or on software or hardware as a service. Instead of acquiring their own computers and software, today companies purchase such resources as a service, with great economic and operational advantages. To this end it they rely on telecommunications networks that enable the connection between their business units and service providers for data processing, software ac‑ cess, databases and safe data storage resources, etc. It is in this segment that the new competitive operators are able to offer services with quality and price differentials that represent better value options compared with those of the incumbent operators. The new computing technologies offer significant leverage to increase the competitiveness of companies. But hese technology breakthroughs pre‑ suppose secure and economic access to cutting-edge telecommunications services. Without such services, companies in any sector are unable to op‑ erate, since nowadays even for issuing simple invoices it is necessary to be connected to finance secretariats, otherwise their access to resources such as e-commerce, among others, becomes severed. The new operators have significantly incresased the supply of quality ser‑ vices for the corporate market, enabling investments and the generation of skilled jobs.

REGULATION AND BENEFITS OF COMPETITION In this section we summarize the essence of economic regulation neces‑ sary to make competition feasible in a sector characterized by natural monop‑ oly and significant barriers to entrance. Competition, according to modern economic literature, is an instrument that leads to efficiency in resource allocation, investment and production. Socially undesirable effects, such as price and quality abuse or unsatisfactory

290

joão moura

service provision, may be mitigated by the simple possibility of switching sup‑ pliers by the consumer, as long as the transaction costs are not exorbitant. When free market forces are not enough to allow choice of providers, reg‑ ulation is necessary until a competitive environment is created, with alloca‑ tion and dynamic effects that favor market development. Reducing barriers to entry and allowing newcomers to bring alternative offers to the market is one of the instruments that regulators have to stimulate the emergence of com‑ petition. In a free competitive trade environment not only does the consumer win, as she has the power to choose – so do established companies that then have incentives to improve their offers, to innovate and, if not to grow, at least to stay in the market. Introducing competitive markets – where feasible – was one of the main objectives of the institutional reforms in Brazil’s telecommu‑ nications sector introduced in the 1990s, reforms that also sought to prevent the socially undesirable formation of private monopolies. As a consequence, the regulatory model itself foresaw some basic ground rules to safeguard free competition: (i) compulsory interconnection of the networks that provide services for the general public; (ii) non-discrimina‑ tory access for clients to service providers that compete among themselves; (iii) a non-discriminatory numbering plan; (iv) access to the open network by competitors under adequate conditions; (v) elimination of cross-subsidies between services; (vi) regulation of prices charged by dominant operators; (vii) non-discriminatory access to rights of way; and (viii) conflict resolution be‑ tween operators through the regulatory agency. It was expected that the institutionalization of these rules would prevent companies with infrastructure networks granted through administrative agreements with the Federal Government from maintaining a privileged competitive position, abusing the rights of such position, preventing the de‑ velopment of competition through the entry of new providers, and limiting achievemnt of the public objectives established in the model.

The basis for regulation of firms with Significant Market Power Market power is the capacity of a firm to set prices above its costs (includ‑ ing the opportunity cost of capital), that is, earning “economic rents”. In all theoretical models of perfect competition, companies reach the sales level that maximizes the consumer’s wellbeing at a price equal to the marginal cost, leading these companies to produce the optimal amount of the product. In


291

these circumstances the factors of production are adequately remunerated, but it is not possible to obtain excess profits (DELORME PRADO, 2011). Most companies have a certain amount of market power. But it is im‑ portant that in a competitive environment this market power always be tem‑ porary or renewable through innovation. In a dynamic context the most real‑ istic reference for the mechanism of competition was proposed by Austrian economist Joseph Schumpeter (1961). He maintained that businessmen respond to competitive pressure with innovation as competition leads to continuous reductions in their profit margins. Innovation allows entrepre‑ neurs to benefit from a monopoly power for a time, that is, it allows them, by means of this power, to earn economic rents. As competitors respond to this dominating position in the same way, that is, through innovations, the companies are prompted to constantly enhance their efficiency in order to survive in the market. Therefore, market power itself is not an unlawful competitive act to be punished. What is repressed by ex ante economic regulation and by the Antitrust legislation (ex post regulation) is the abuse of market power. It is this abuse – by means of deception that aim at its continuity – that needs to be punished (DELORME PRADO, 2011), because this negatively impacts the competitive environment and consumer’s wellbeing. The ideal scenario is not a market with perfect competition but a market where competition plays its role of challenging the dominant company (com‑ panies); that is, the companies having market power and capable of exercising it in different degrees. Nevertheless, what is intended with anti-trust legisla‑ tion and ex ante economic regulation is to prevent the perpetuation of this market power through actions derived from the abusive exercise of a domi‑ nant position. Such actions include the practice of predatory pricing, margin squeeze, deterioration of the quality of the services provided to competitors themselves as required by regulation, such as leased circuits and interconnec‑ tion, that affect the quality of the service provided by the competitive operators to their final clients. In short, market power that results from the search for efficiency and that, as a consequence, implies that the company is always more efficient than its competitors, does not harm the wellbeing of consumers – on the contrary, it not only increases such wellbeing, but also fosters the country’s technical progress and economic growth and directly benefits the consumer.

292

joão moura

THE GENERAL PLAN FOR COMPETITION GOALS (PGMC) Anatel’s PGMC aims to curb abusive behavior of economic power, through ex ante regulation, with asymmetric rules being applied to holders of economic power, known as operators with Significant Market Power (SMP), as analyzed in detail in Chapter 4. Below we will discuss international regula‑ tory experiences and the PGMC in Brazil.

Share or duplicate networks? The European experience Competition legislation in the telecommunications sector in Brazil is based on the European model. The first attempts at asymmetrical regu‑ lation attempts based on SMP ocurred in 2005. But only in 2012, with the publication of the PGMC, did Anatel formalize the mindset idealized by the Europeans in a set of rules, the essence of which is summarized as follows. Competition plays a crucial role in the sector but it only reaches consoli‑ dation if regulation strikes a balance related to the huge differential between dominant operators that control legacy networks built more than a century ago and new entrants that invest in innovative networks and services. Regulation is not an end but a means to mitigate disparities and promote competitive dynamics with minimum external interference. Regulation implies constant adjustments due to the evolution of markets and the results attained. It is crucial to evaluate results and compare experiences. There is no merit in sim‑ ply copying models from others, but reflecting on experiences brings inspi‑ rations and increases the chances of correct decisions. Generally speaking, the European model encouraged network and operator sharing, where the incumbent operators were required to lease network assets to the newcomers, their competitors. In this way the regulators sought to facilitate initial invest‑ ments by the new entrants so that they that did not need create their own ex‑ tensive networks to enter the market. Consequently the provision of new ser‑ vices increased. Allowing new entrants access to the networks of incumbent operators is regulated as regards pricing and commercial conditions. Pricing, based on cost models seeks to avoid cross-subsidization between operators and to keep incentives for investments, not only by incumbent operators but also by the new ones (ALLEN, 2015). In Brazil, Anatel followed the same approach, however it acted much less rigorously in the regulation of prices and supply conditions. Consequently we did not have the same success in providing incentives for investment and expansion.


293

In the US the situation was different and regulation did not create sub‑ stantial sharing obligations, possibly on the assumption that there were al‑ ready two alternative networks (phone and cable TV), which would suffice to enable competition. The differences between the situation in the European and in the American markets, where the concentration of large groups prevails, are illustrated in an article from the BBC (GEOGHEGAN, 2013) that compares services offered and points out that prices in San Francisco, New York and Washington DC are, respectively, $99, $70 and $68, compared to $38 in London, $35 in Paris and $15 in Seoul, for the most common speeds for broadband.5 Competition is a crucial factor for stimulating investments not only by alternative operators but also by the incumbents, that tend to invest only in response to the competitive pressure of new operators. In Europe, access reg‑ ulation was (and still is) decisive for promoting investments by alternative op‑ erators, generating provison expansion, increasing average broadband speed and forcing price reduction. In Brazil, two ways of regulating access, local loop unbundling and the market for “industrial exploitation of dedicated lines (Exploração Industrial de Linha Dedicada – EILD) – that is leasing of lines for service provision between operators) – have never been effective. On the con‑ trary, they are conflict-ridden and challenged by the dominant companies, as Anatel itself recognized on many occasions. If on one hand Anatel supports the principles followed by the European regulators, on the other it is reticent in the practical rollout of access regula‑ tion, arguing that the obligation to open legacy networks favors opportunists that neither invest nor innovate. In practice, Anatel seems to fear the exten‑ sive application of the concepts of “investment ladder” (CAVE, 2006), which is one of the conceptual pillars of European regulation, and therefore seems to encourage opportunists and undermine the operators with SMP. The core idea of “investment ladder” is that the combination of own net‑ works with leased networks allows the newcomer to control the rate of con‑ struction and to mitigate market risks, making investment more efficient. Moreover, Anatel seems to disagree with the fact that building own network is the only way either to innovate or to invest. Adding value by means of in‑ novative services provided on networks leased from third parties, based on prices that are defined by fair economic criteria, that is, prices that allow a due 5

Available on: . Accessed: March 31st., 2016.

294

joão moura

return on investment, is not an opportunistic behavior. On the contrary, it allows increasing the supply and the diversity of alternatives for the client and also stimulates other investments in value-added services and dynamizes the market. That’s the value of competition.

Evaluation of the PGMC Despite the importance rules approved regulating competition in Brazil, their rollout has been hesitatiing and still has not brought the expected positive results, though more than two years have elapsed since the PGMC was issued. Access regulation (with multiple technologies) is a key element of the PGMC for enabling the expansion of alternative networks. Without the pos‑ sibility of leasing network complements, the investments by competitive op‑ erators become less attractive in terms of risk and return and are made in a slower pace. As pointed out previously, access regulation in Brazil was not implemented in an effective manner, neither immediately after privatization nor recently with the PGMC. With the prices and conditions of the offers ap‑ proved by Anatel, the new operators have preferred to build their own ones, even if that leads to slower growth. In Brazil, access regulation is even more crucial as an investment-curbing element because the regulations for the use of poles, ducts, and rights of way are incipient and that makes network construction by competitive operators even more burdensome and difficult.6 In short, in Europe regulators sought to promote competition in services and all operators had access to the legacy networks built in the past. Therefore it was possible to stimulate new complementary investments or in more modern stretches of networks combined with old networks. The results were positive if we take into account indicators such as coverage, penetration, prices and broad‑ band speed. In the US, the model prioritized competition among networks, based on two large existing infrastructures, for telephony and cable TV. In Brazil the PGMC adopted the European conceptual framework, how‑ ever its implementation was fragile and imposed extremely light obligationos on the old monopolists. As a consequence, the expansion of broadband access in Brazil grows sluggishly and below potential.

The PGMC and the movements of business concentration 6

This issue is tackled in detail in the Chapters 4 and 11 of this book.


295

In capital-intensive sectors the search for scale is a natural strategy and therefore mergers and acquisitions are common. In this aspect Brazil is no exception. Competitive operators, when they reach a certain size, become the target of the large operators. The logic is to increase scale and remove from the market an aggressive competitor and therefore decrease competi‑ tive tension. There is advantage in cost reduction, in opportunities to ratio‑ nalize investments and also there is better flexibility in the market in terms of pricing policies. The consolidation process is healthy for it offers entre‑ preneurs the opportunity to monetize their investments. On the other hand, it is the responsibility of the anti-trust bodies and of the regulators to act in such a way so as to avoid excessive concentration that can become harmful for the market in general. The PGMC provides the necessary framework for the analysis of the con‑ centration acts and it could avoid damages to the competitive environment. Nevertheless, as with every model, it all depends on its practical application and on the regulator’s judgment. In Brazil, Anatel as well as Cade (defense of competition) have been flexible, approving mergers operations with relative ease and with little or no interference. On the contrary they even interpret reg‑ ulations in a more flexible manner to facilitate transactions, as was the case of the merger of Oi with Brasil Telecom. US authorities have been stricter in their defense of competition, blocking transactions or imposing substantial compensation obligations, including massive investment commitments.7

CONCLUSIONS: THE GOVERNMENT’S ROLE AND THE CONTRIBUTION OF THE LARGE AND MEDIUM OPERATORS Telecommunications in Brazil have progressed in a consistent manner and nowadays the country is in better position in this sector – in international comparisons – than in other infrastructure segments, such as roads, ports, airports and urban transport. The model designed for fixed telephony expansion, the great aspiration of the population until the 1990s, now needs massive reconfiguration to allow ex‑ pansion of high-quality infrastructure and greater access to broadband Internet services. Only then will the digital economy be able to advance and contribute to the Brazil’s competitiveness and to the country’s economic and social growth. 7

For instance, the AT&T & DirecTV case, approved with massive investment liability in new generation networks, FCC-15-94 A 1 – July 28th, 2015.

296

joão moura

The digital economy serves as a fundamental enabler for other sectors that need to move forward in the country. Education, public services, industry, pro‑ fessional services and many other sectors today require access to 21st century infrastructure and next generation networks to thrive. New business models created by the Internet cannot move forward in Brazil without this resource. Developing this infrastructure requires the mobilization of society, busi‑ nesses and the government. There is room enough for many agents and mul‑ tiple possibilities action consistent with these goals. Below some suggestions are provided.

Governments Public policies in each of the three levels of Brazil’s federal system (mu‑ nicipal, state and federal) play a paramount role and may contribute a lot more than they do today. Unlike other sectors in Brazil, telecommunications has not received public investment since privatization.8 Tax and financial incentives have been generously allocated to traditional industries such as automobile pro‑ duction, meatpacking and household appliance manufacture. Meanwhile the telecommunications sector has been a strong contributor to government reve‑ nues. The gross operational revenue of telecommunications service providers amounted to 4,2% of the GDP in 2014 and they contributed R$ 60.1 billion in taxes, that is 43% of their net operational revenue (TELEBRASIL, 2015, p. 7). In February 2016 a minimum of 30% to 42% of household telecommunica‑ tions bills in February 2016 corresponds to taxes, depending on the state.9 That places a heavy burden on the consumer, inhibits usage and reduces providers’ in‑ vestment capacity. At all levels of the federation the priority has been to collect as much revenue as possible as quickly as possible. For instance frequency auctions seek to maximize revenues to the federal government, “contributions to sectoral development funds are are channeled to the national treasury and the the states’ value added tax (Imposto sobre Circulação de Mercadorias e Prestação de Serviços – ICMS) has extremelty high rates, and they were raised even higher at the end of 2015 to cover deficits in public accounts. It is hard to conciliate official declarations on the importance of telecommunications and digital inclusion for the country with that the exceptionally high tax burden imposed on telecommunications 8 In 2010, the federal government decided to reactivate Telebras and it is making some investments in networks, sattelites and undersea cables. Nevertheless it is still not clear how it will position itself in the market. 9

Available on: . Accessed on: Feb.11th, 2016.


297

services, including broadband.10 If the governments (the three levels) wanted to do something, simply reducing the tax burden would contribute a great deal and would create much better economic environment in the future.

Regulatory agencies and Anatel The regulatory agency model, as explained at the beginning of this chap‑ ter, was created as part of the reform and privatization of the infrastructure sector. Though well structured, the model did not come in for much recogni‑ tion as regards its importance for society. On one hand, the executive branch of government at times seems not to believe in the model and interfers in the work of the agencies, directly, by meddling in their operations, as well as indirectly, by appointing politicians to technical positions, leaving positions vacant for long periods (impeding the agencies’ work) and even freezing part of their budgets, thereby limiting their ability to carry out their duties. As state bodies and with a long-term perspective, the agencies need clear planning and competent performance. Sectoral policies, competition, invest‑ ment incentives, service quality, security and client service are all relevant issues. Without a strategic focus and coherent performance no results can be attained. Consequently the agencies’ agenda becomes dominated by politics due to external pressure on the part of interest groups, the press, etc. and that undermines effectiveness. Regulation becomes but one more burocratic activ‑ ity, without generating the value society could receive. Anatel is no exception. It has suffered the same difficulties as the other agencies, with more or less intensity, depending on the whim of the moment. Moreover it suffers the consequences of its own choices. As it does not have a clear plan of action, with consistent accountability for the results of its actions, its institutional reputation remains poor and it lacks due support from other state and federal entities and from public opinion.

The large companies The sector has developed over the last two decades with a focus on fixed telephony and later on on mobile service. In both cases it achieved almost universal access to these services for the population, though the quality and prices of the services still leave much to be desired.

10 The issue on telecommunications services taxation is tackled in detail in this book in Chapter 8.

298

joão moura

In broadband achievements have been more modest as demonstrated in Chapter 1. International statistics show that Brazil still has a long path ahead, es‑ pecially in terms of fixed broadband provision and new generation networks. In mobile broadband, service penetration advances quickly with the dissemination of smartphones. Service usage and data volume circulation are low and prices high. Despite the merits of mobile broadband, high-reliability and high-capacity fixed line networks are essential, mainly to meet the needs of the corporate mar‑ ket. Small and medium companies need more and better services to develop their businesses and boost their productivity and competitiveness. In Brazil the advancement of fixed broadband has been hampered by two factors: 1. The lack of effective access regulation for unbundling last mile networks controlled by the fixed telephony operating companies. The unbundling process (as the process of disaggregating telephony operator networks is called) was the means used, especially in Europe, to facilitate the en‑ trance of new competitors in the market. Consequently a vicious cycle emerged, where competition encourages defensive investment by the incumbents. In Brazil, unbundling is still at an early stage. 2. Delays in the legislation subscription or “cable” TV. For about nine years no license for cable TV networks was issued in Brazil. The cable networks, with technological convergence, proved themselves to be extremely ef‑ fective in broadband provision. Only after the entry into force of the the law of conditioned access or access by subscription (Lei de Acesso Condicionado ou por Assinatura – SeAC, Law n. 12,485/2011), that allowed the issuing of new licenses, did the cable networks resume expansion and nowadays they account for a significant portion of fixed broadband service provision in the country.

The competitive operators: two business models for two market segments In this scenario of restricted provision and growing demand, room has been created for two new business models that nowadays progress at a swift pace. The first focuses on the corporate market and it is a priority for most members of TelComp. This segment has shown dynamism and growth capacity despite the lack of effective regulatory incentives. Even duriing economic downturns,


299

telecommunications services for the corporate market are extremely important and have a great potential to increase business productivity and job creation. The second focuses on the residential market and brings together small‑ er providers, most of them belonging to organizations such as the Brazilian Association of Internet and Telecommunications Providers (Associação Brasileira de Provedores de Internet e Telecomunicações – Abrint) and the Brazilian Internet Association (Associação Brasileira de Internet – Abranet). These providers are able to offer services in regions that are not prioritized by traditional operators.

Telecommunications in Internet economy The Internet has disrupted traditional sectors, such as music, video, book pub‑ lishing, retail trade, hotels, taxis and advertising, and created digital giants that vie for new market segments, such as search, social networks and cloud computing. The telecommunications sector did not go unharmed. In the traditional model telecommunicatons operators controlled their own networks and the services they provided. In the IP world reality is very different. App and content creators have access to telecommunications networks without the intervention of operators. That makes room for the release of apps, including those that com‑ pete with and directly replace basic telecommunications services themselves. This movement requires telecommunications operators to meet the chal‑ lenge of reinventing themselves in the Internet’s digital economy. As we are confident that telecommunications networks will not disappear, but rather grow rapidly, it seems evident that a solution will emerge and tele‑ communications operators will enter this digital environment and keep on providing services of high value for society.

Critical aspects to enhance public policies In order to advance in the digital economy the country needs to count on more and better telecommunications networks and, to that end, well-co‑ ordinated public policies are necessary. Such policies should include the following aspects:

Taxation Massive digital inclusion in Brazil is impossible with the present tax burden imposed on telecommunications services. If the country believes in

300

joão moura

digital economy and we want to popularize broadband, we will not reach that with 50% taxation on every telecommunications service bill.

Regulation The PGMC is an adequate benchmark framework for stimulating compe‑ tition, and also innovation, quality and prices. It is crucial to evaluate results on a continuous basis and adjust its parameters for the Plan to work and play the role for which it was designed.

Investment Costs Investing in Brazil is a lot more costly than in other countries. Bureaucracy at all levels, lack of planning integrated with other infrastructure investments, and insufficient resource sharing, among others factors, result in duplication of expenses and compromise investment productivity.

REFERENCES ALLEN, J. The digital single market and telecoms regulation going forward: Report for ECTA – European Competitive Telecommunicatons Association. Analysis Mason, Sept. 18th, 2015. Available on on: . Accessed: February 12th, 2016. CAVE, M. Encouraging infrastructure competition via the ladder of investment. In: Telecommunications Policy, n. 30, p. 223-237, 2006. DELORME PRADO, L. C. Contribuições PGMC [PGMC Contributions]. Study financed by TelComp, September, 2011. GEOGHEGAN, T. Why is broadband more expensive in the US? BBC News, Oct 28th, 2013. Available on: . Accessed: February 12th, 2016 SCHUMPETER, J. A. Capitalismo, Socialismo e Democracia [Capitalism, Socialism and Democracy]. Rio de Janeiro: Fundo de Cultura, 1961. TELEBRASIL. O Desempenho do Setor de Telecomunicações no Brasil: séries Temporais 9M15. [The Performance of the Telecommunications Sector in Brazil: 9M15 Time series] Rio de Janeiro, Dec. 2015. Available on: . Accessed: February 7th, 2016

case studies

| 13 |

Public-private partnerships for expanding broadband access: Lessons from the Cinturão Digital do Ceará Network in Brazil Fernando Carvalho [ Universidade Federal do Ceará ] Flavio Feferman [ University of California, Berkeley ] Peter Knight [ Instituto Fernand Braudel de Economia Mundial ] Glenn Woroch [ University of California, Berkeley ]

D

INTRODUCTION

espite evidence that access to fast, affordable digital infrastructure contributes to economic development and social inclusion, the eco‑ nomic realities of broadband projects can limit their deployment. The high cost of reaching remote communities and the paltry revenues from low-income inhabitants often make these projects unprofitable. In response, publicly-owned broadband networks have been built where for-profit alter‑ natives are lacking. In other cases public-private collaborations have been formed with the goal of meeting these needs by capturing the benefits of both public and private ownership. A unique collaboration of public and private interests was at the heart of the Ceará Digital Beltway (Cinturão Digital do Ceará, “CDC”) project that is the subject of this chapter. The CDC is a fiber network with wireless ex‑ tensions that delivers high-speed data transmission throughout the Brazilian state of Ceará. This project was undertaken by the State of Ceará’s ICT com‑ pany Empresa de Tecnologia da Informação do Ceará (Etice) which specified the terms and conditions for collaboration with other public organizations and with private sector partners. The CDC model innovates on the convention‑ al public-private partnership (PPP) to conform to the economic and political conditions that prevail in the target markets.

304

f. carvalho • f. ferferman • p. knight • g. woroch

The chapter begins by examining the history and economics of the CDC in the broader context of Internet development in Brazil and worldwide. We show how the state of Ceará brought high-speed data services to underserved areas by combining public financing and ownership with the profit incen‑ tives of private enterprise. Importantly, over time, Etice adapted its model by heeding the feedback of its constituents and partners and responding to market conditions. Specifically, in the second part of the chapter, we describe the challenges faced during “Phase III” of the CDC project, when the state of Ceará structured a concession/PPP to lease excess dark fibers to private sector companies, and how the terms of the concession were modified to promote private sector participation. We discuss the “theory of change” behind the CDC, examine the complementary roles of public and private investments for broadband expansion, and illustrate how government investments can (par‑ adoxically) promote private sector participation and competition. Finally, we address whether the unique CDC business model could provide an effective solution to the deployment of broadband services in other parts of Brazil as well as in other developing countries.

THE CDC IN THE BROADER CONTEXT OF INTERNET DEVELOPMENT IN BRAZIL AND WORLDWIDE1 Beginning in the late 1980s, inspired by the development of the Internet in the United States and Europe and by collaborations with some of its early pioneers, Brazilian academic researchers laid the groundwork for connecting local networks to the global Internet. Together with government supporters, they launched the National Research Network, now National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP), modeled af‑ ter NSFNet in the United States. The RNP was first connected to the glob‑ al Internet during the United Nations Conference on the Environment and Development in May 1992. The link was established by the federal state tele‑ communications company, Embratel. In that year the RNP connected eleven Brazilian cities at speeds of 64 Kbps or 9.6 Kbps. It was not until 1995, however, that commercial use of the Internet arrived in Brazil, with Embratel providing the backbone network and a multitude of small Internet Service Providers (ISPs) providing last-mile access to busi‑ ness and residential customers. To this day small and medium ISPs play an 1

This and the following sections are based on Knight (2014, Chapter 2).

Public-private partnerships for expanding broadband access

305

important role in the expansion of the Brazilian Internet. In that same year, the Brazilian Internet Steering Committee (Comitê Gestor da Internet no Brasil, “CGI.br”), a multi-stakeholder body performing functions similar to ICANN (Internet Corporation for Assigned Names and Numbers) was launched (CGI predated ICANN by three years). The privatization of the Brazilian telecom‑ munications industry took place in 1998, leading to a rapid expansion of the commercial Internet, supervised by CGI.br. Backbone services were provided by commercial carriers that gave priority to major urban centers, especially their higher-income areas. The small and medium ISPs established service in less profitable markets and relied on the four major telecommunications companies (“telcos”) for access to the Internet. In some markets, the major telcos gained signficant market power (as discussed in Chapter 4), including in Ceará where a single company dominated the market.

Brazilian Policies for Broadband Expansion In 1997, Brazil introduced the basic regulatory framework that governs the Brazilian telecommunications industry, the General Telecommunications Law of 1997. This law also created the National Telecommunications Agency (Agência Nacional de Telecomunicações – Anatel) a regulatory entity charged with protecting the interests of the State and citizens, encouraging competi‑ tion, universalizing telecommunications services, and modernizing telecom‑ munications infrastructure. The Telecommunications Law and Anatel provid‑ ed a framework for subsequent investments by the private and public sectors. Two other initiatives contributed to broadband expansion in Brazil. In 2007 RNP launched a program to create metropolitan networks in state cap‑ itals and other major cities. A few years later, in 2010, the Brazilian federal government launched the National Broadband Program (Plano Nacional de Banda Larga – PNBL) to address deficiencies in Internet access. The principal instrument for implementing the PNBL has been Telebras, a federal telecom‑ munications holding company that was put into a state of hibernation during the privatization period and was re-activated in 2010 as an operating company for this purpose. Funding of Telebras to achieve PNBL objectives has been modest, in large part due to competing demands on limited federal budgets. However some progress has been made in creating a national backbone. In parallel, the RNP continued to expand its inter-urban backbone to connect research and higher education institutions dispersed throughout the country. By 2015 RNP had

306


points-of-presence (“POPs”) in all 26 states and the Federal District. The newest development involving RNP is a program called New Paths (Veredas Novas) to bring high-speed broadband to research and higher education insti‑ tutions outside the capitals of the 26 Brazilian states. (see Chapter 9) Despite these public, private, and academic initiatives, in 2014 only 25% of households in Brazil had fixed broadband Internet connections of 2 Gbps or higher (calculated based on CGI.BR, 2015, Tables A4 and A6, pp. 322 and 324). Digital inclusion of individuals (defined as access to the Internet in the three months prior to surveys) reached 55% in 2014, but remained much low‑ er in rural areas at 29%. A large gap in Internet use persisted between the richest segment of the population (Class A, at a 96% rate) and the poorest segments (Classes D and E, at just 21%) (CGI.br, 2015, Table C2, p. 335).2 Broadband service in Brazil remains slow, expensive, and often unreliable (see Chapter 1). The state of Ceará is located in the relatively poor Northeast region of Brazil. The state’s patterns of Internet access are similar to those in the rest of the country, with deficiencies in service availability and quality, as well as gaps between urban and rural areas and across socioeconomic groups. As described below, the CDC initiative was introduced to expand broadband access throughout the state, including the capital city of Fortaleza, smaller towns, and rural areas. In the following sections, we describe the evolution of the CDC, its innovative public-private partnership (PPP) model, and how the CDC network is promoting economic development and innovation through‑ out the state.

2

CETIC.br, accessed February 7, 2016.


307

Figure 13.1. Location of the State of Ceará, Brazil

Source: http://www.luventicus.org/maps/brazil/ceara.html (Accessed September 2, 2016)

HISTORY OF ETICE AND THE EVOLUTION OF THE CDC NETWORK The CDC is a pioneering project to promote statewide connectivity and digital inclusion, and is considered Ceará’s highest profile ICT initiative. The CDC is Brazil’s largest and fastest public broadband network, reaching almost 86% of the urban population of the state. By the end of 2015, the CDC covered nearly 100 of the state’s 184 municipalities. The CDC was developed by Etice over three phases, as illustrated in Figure 13.2. The first phase involved the initial planning and proof of concept for the CDC: in partnership with the RNP academic network, Etice developed the Gigafor network in the capital city of Fortaleza. The second phase involved the construction and initial operation of the network, which included partner‑ ships with the state electricity company, regional ISPs, and numerous state and municipal government agencies. The third phase involved a more formal

308


concession model (PPP) for leasing excess dark fiber to private sector com‑ panies. Thus, each phase of the project included different types of public, private, and academic partnerships.

Figure 13.2. CDC Timeline ETICE becomes state CIO, begins CDC planning and proof concept (Gigafor).

2007

CDC construction begins

2008

Phase I: Planning and Proof of Concept Concept (Gigafor)

2009

First concesssion Edital

CDC construction completed

2010

2011

Phase II: Initial Construction and Expansion

2012

Second concesssion Edital

2013

2014

Concession contract signed

2015

Phase III: Private Sector Concession and Further Expansion Fonte: Elaboração dos autores.

This section discusses the context and business model for the initial devel‑ opment and expansion of the CDC during Phases I and II, describes the net‑ work infrastructure, and discusses the results and achievements of the CDC during its first two phases. Subsequently, the chapter will describe the third phase of development in more detail, focusing on the PPP concession model for leasing dark fiber to private sector companies.

Background on Etice Etice was created in 2001, with the closure of its predecessor, Seproce (Serviço de Processamento de Dados do Estado do Ceará), the state-owned data processing company. Etice was established as an autonomous public compa‑ ny linked to the Ceará State Planning and Management Secretariat (SEPLAG). Starting in 2007 Etice took on a much broader role, becoming responsible for both ICT services and strategic planning for the State of Ceará, with its president acting as a de facto state Chief Information Officer (CIO). The com‑ pany has control of its own assets, and benefits from administrative, financial and technical autonomy in its activities. When a new state government took


309

office in 2007, a plan was developed to build a modern telecommunications infrastructure in Ceará. The central idea was to build a fiber optic network that reached all areas of the state and to provide Etice with the human and financial resources to operate this network.

CDC Objectives The CDC project had two principal objectives. The first objective was to reduce the operating expenditures (“OPEX”) of the state of Ceará with con‑ nectivity services. At the time, the state was paying high rates to a regional, monopolistic operator, totaling R$29 million per year (about US$15 million in 2007), for the use of an obsolete network that was built to support fixed line telephony. Etice’s initial investment in the CDC was R$68 million (about US$35 million at the time), including the deployment of fiber and equipment. The goal was to recover this investment by reducing the telecommunications costs of the government, especially in towns with more than 50,000 inhab‑ itants. By the beginning 2015, the operational cost of the network had been reduced to only R$7 million per year (US$2.6 million at the time), with sig‑ nificantly improved quality. By December of 2015, the CDC network covered 86% of the municipalities in the state, and Etice paid the main regional oper‑ ator a much lower fee of R$6 million/year to reach areas without CDC cover‑ age. Considering the current operating costs and remaining operator fee, the CDC reduced annual telecommunications costs by 55% for the state of Ceará, by the end of 2015. The second principal objective of the CDC was to provide universal broad‑ band coverage throughout Ceará, including the capital city of Fortaleza, small‑ er towns and rural areas. In 2015, public entities, including municipal gov‑ ernments, were served with high-speed connectivity and paid a monthly rate based on their amount of inbound and outbound traffic. Public schools, cus‑ toms offices, hospitals, police stations and many types of public services run their applications over this infrastructure. These applications include services such as telemedicine, distance education, security camera networks, remote scanners for cargoes, and remote management of transportation infrastruc‑ ture. The network also supports the administrative operations of government agencies in the state, as well as numerous innovative technology firms in Ceará, which also utilize the CDC.

310


The CDC Business Model The CDC aimed to achieve long-term financial sustainability using a busi‑ ness model that includes (i) usage-based fees for state government clients, (ii) sale of network services to municipalities and local populations at cost, with involvement of regional ISPs, and (iii) lease of network capacity to private sector concessionaires to cover Etice’s operational costs and maintenance ex‑ penses with the CDC network (Phase III). At the end of 2015, the CDC supplied Internet access to 1.5 million users, linking more than 2,000 points throughout the state, at speeds up to 1 Gbps. Monthly fees are based on usage. For each Gigabyte downloaded per month, in 2014 Etice charged R$7.10 (about US$3.00 at the exchange rate in mid2014). This monthly fee helps to cover the operational costs of the installed network, including an overhead cost of about 15%. It is important to note that any user wanting to expand the fiber optic network can deploy new cable, pro‑ vided these conform to Etice’s technical standards. (This model is similar to the one utilized for the expansion of electricity distribution networks.) When the cable becomes operational, the user begins receiving services from Etice. By the end of 2015, the network initiated by Etice had been expanded by about 25% through this type of arrangement. In addition to state government entities, municipal governments also participate in the CDC through public requests for proposal (RFPs) to re‑ ceive, at cost, the transfer of data from Fortaleza. By December of 2015, forty-five municipalities were participating in this program. It is important to note that, prior to the CDC, the cost of Internet in the interior of the state was very high, given the scarcity of fiber connections.3 By 2015 municipali‑ ties received Internet access at a cost reduction of more than 90% per Mbps, as compared to costs prior to the CDC. Each municipality hires ISPs to deliv‑ er Internet access in Fortaleza, at the Etice headquarters. Etice then provides transport at up to 200 Mbps to the municipality, which then distributes this bandwidth locally through local ISPs. As the municipal governments on average need only 20 Mbps, local ISPs deploy the excess bandwidth (about 180 Mbps) in services to the general public. Etice plans to expand this type of initiative, which supports municipal governments, ISPs in Fortaleza, and local ISP firms at the municipal level. The model also provides local citizens 3

Prior to the CDC, municipal contracts with providers ranged from R$9,000-R$16,000 for 5 full Mbps; in December 2015 the CDC delivered 200 full Mbps for R$4,000.


311

with high-quality broadband at attractive prices utilizing the excess capacity of the network. Municipal governments provide a matching contribution: they ensure that all schools within the municipality and at least two public squares have access to the Internet via the CDC (local ISPs provide service to the public squares). The academic community has also played a critical role in the devel‑ opment of the CDC. Prior to the CDC launch, Etice conducted a “proof of concept” with RNP to validate the operational and economic viabili‑ ty of the project at a smaller scale. At the beginning of 2007, RNP had built a fiber optic ring in the city of Fortaleza to connect universities and research centers in the metropolitan area. However this infrastructure was not operational because there was no maintenance contract for the network. In exchange for the use of two pairs of fiber for state adminis‑ tration, Etice proposed to maintain the network, through an outsourcing contract. Subsequently, Etice utilized these two pairs of fiber to connect 42 public institutions in Fortaleza through its own network utilizing MetroGigabit Ethernet technology. Etice invested R$1.1 million (US$570,000 at the time), including equipment and additional fiber, with a network speed of 60 Mbps for each point. Etice recovered its investment costs in only 90 days. This initial proof of concept with RNP resulted in the creation Gigafor, a high-speed fiber optic network for the city of Fortaleza, which is now part of the CDC. Gigafor greatly increased the reach of the initial RNP metropolitan network in Fortaleza. After the CDC was deployed, the cooperation with RNP to reach the aca‑ demic community was extended to include the entire state of Ceará. RNP now deploys necessary fiber to expand its network, and Etice manages the network, which reached 45 universities and research institutions in the state connected through a 1 Gbps Metro-Ethernet, by 2015. This experience with RNP served to validate the technology, and provide a template for subsequent agreements with private sector partners. A similar swap arrangement was executed during the initial develop‑ ment of the CDC. Through a partnership with the private energy utility company Enel, that owns the state electrical utility company Companhia Energética do Ceará SA (Coelce), Etice was able to deploy the backbone network at a significantly lower capital outlay than would a purely private network. Etice signed a contract to use Coelce’s electrical poles and sub‑ stations. In compensation for this access, the power company gained the

312


right to use two pairs of fiber in the CDC network. Later, the same formu‑ la was extended to all fiber deployed through Coelce’s poles and towers. Today Coelce utilizes this fiber to control its stations and to expand its smart grid. By swapping access to each other’s infrastructure, Etice and Coelce realized significant cost reductions in the provision of both broad‑ band and electric power services throughout the state. During Phase II, other private firms became increasingly involved in the use and maintenance of the backbone network. Etice negotiated access to the network for a variety of private firms, and retained 8 to 10 fiber pairs, each with a 10 Gigabit transfer capacity available for sharing with companies in future public-private partnerships during Phase III. During the third phase of the project, to be discussed in more detail further below, part of the network capacity was leased under concession agreements to private sector operators. These operators utilize the CDC to provide con‑ nectivity and value added services, and plan to expand the network based on their own business objectives. In parallel, during the third phase, the state of Ceará is expanding the CDC network to new areas of the state, and particularly to areas where private sector operators lack financial incen‑ tives to provide connectivity services. Hence, Phase III of the CDC utilizes a hybrid business model, combining the profit incentives of the private sector with the public sector mission to provide universal broadband ac‑ cess in the state. The public-private model has been well received across industry seg‑ ments. The state investment in the CDC network reduces capital outlays for private firms seeking to expand their services throughout the state. Current and potential future users of the network include large telcos (for Internet service and mobile network backhaul), regional ISPs, cable tele‑ vision companies, fiber-to-home providers and rural telephony providers. Through this PPP approach, Etice seeks to ensure that its investments stimulate private ICT industry development while lowering costs for deliv‑ ery of public services.

CDC Deployment and Technology In May 2008, after the proof of concept with RNP and the swap agree‑ ment with Coelce, Etice began development of the CDC network. Etice con‑ ducted a public bid for the deployment of the fiber optic network, and the con‑ struction of the network began in January 2009 with the company Schahin


313

Engineering. The network was completed in September 2010, on time and on budget. In 2011, the state of Ceará passed Law 15,018 which designated Etice as the manager of the CDC. The CDC initially consisted of about 3,000 kilometers of fiber optic ca‑ bles connecting the 50 most populous cities in the state. This connectivity was augmented by high-capacity wireless WiMAX (IEEE 802.16) radios for delivery of broadband connections beyond the reach of the fiber optics ca‑ bles. In addition, lower capacity Wi-Fi radios were often deployed for local‑ ized connectivity, including free community Wi-Fi in the central plazas of several cities in the state. The core network consists of 24 fibers in the main ring, linear extensions that deploy 24 fibers to larger cities and 12 fibers to smaller cities, plus lastmile deployments. In the main ring, extensions, and last mile, some of the fi‑ bers are utilized by Etice and other partners (Coelce, RNP), while other fibers remain unlit and are available for future concessions or swaps. The service provided by Etice is based on the technology “Ethernet over DWDM” (Dense Wavelength Division Multiplexing) (SAENGUDOMLERT, MODIANO, e GALLAGER, 2006). There are 65 stations separated by an average distance of 40 kilometers. These stations regenerate the signal and have IP switches. The stations are strategically located in proximity to the 50 largest cities and towns in the state, and the optical ring provides redundancy for the network. The installed capacity is 20 Gbps between stations. Figure 13.3 illustrates the routes covered by the CDC in 2014.

314


Figure 13.3. CDC state network in 2014

Source: Etice


315

In densely populated urban areas, Metro-Ethernet technology is used, with speeds of 1 and 10 Gbps. In the capital Fortaleza, the Gigafor network uses GPON (Gigabit Passive Optical Network) technology to reach 811 sub‑ scriber points, including schools. It is worth noting that in 55 municipalities there are towers with point-multipoint radio, using WiMAX (IEEE 802.16) in the 3.5 and 4.9 Ghz frequencies, with the latter being used only for pub‑ lic safety and rescue services. All the Personal Mobile Radio (PMR) ser‑ vices of the state police utilize the fiber optic infrastructure of the CDC, through a technology called Terrestrial Trunked Radio (Tetrapol) over IP. Five thousand portable and vehicle-based terminals have been installed in the Tetrapol network.

THE “THEORY OF CHANGE” FOR THE CDC AND THE IMPACT OF EXPANDING BROADBAND ACCESS The CDC planners understood that both the public and private sectors had distinct advantages in promoting broadband expansion and were com‑ plementary with one another. As indicated in Figure 13.4, by 2007, when the CDC project was initially conceived, the private sector already provided broadband services to high income, low cost areas of Northeast Brazil – pri‑ marily in state capitals such as Fortaleza. In many cases, those capital cities were served by monopolistic incumbents. At the other end of the spectrum, Etice realized that public investments in infrastructure (or other new public policies) would be required to reach lower income areas with high service costs, such as small towns and rural areas. In between these two zones, Etice believed that the “access frontier” could be expanded through public-private partnerships, and specifically through concessions during Phase III of the project. It was envisioned that the CDC concessions would stimulate com‑ petition and market entry of smaller players, which would utilize the excess capacity of the network to expand access. Thus, Figure 13.4 highlights that each ownership model has a comparative advantage depending on the mar‑ ket served.

316


Low cost areas (Class A, cities)

Remote areas w/ high cost (Periphery & interior)

Figure 13.4. Strategy to expand access through both public investment and private participation

Low income, interior (30%) Access Frontier

Range of fixed network in 2007 (monopoly) Best areas of capital cities (10% of population) High Income $$$$

Role of Government

Role of private sector, supported by government actions

Competition

Low income $ Source: Adapted from World Bank (2011), Figure 6.2, p. 157.

A number of approaches have been utilized to promote broadband expan‑ sion in the “access frontier” zone illustrated in Figure 13.4. These approaches include: • Reverse subsidy auctions, where the public sector offers subsidies to encourage the private sector to invest in areas that are commercially non-viable. In this type of auction, the winning bidder is the company that requests the lowest subsidy amount necessary to execute the project and deliver the required services. This approach was utilized in coun‑ tries such as Chile, Peru, India, Nepal, and Peru (WALLSTEN, 2008). Another example of this is the Universal Communications Service Access Fund (UCSAF) in Tanzania, funded by the World Bank. • Shared ownership arrangements, such as the Alberta SuperNet case, discussed further below; • Different flavors of public-private-partnerships under a continuum of potential arrangements illustrated in Figure 13.5 (KWAK, CHIH, and IBBS, 2009). A well-known example in Brazil is the PPP for the de‑ velopment of the Banco do Brasil and Caixa Econômica Federal data center, which followed a Build-Operate-Transfer (BOT) model.


317

Figure 13.5. Continuum of Types of PPP Design-BuildOperate (DBO)

Build-OperateTransfer (BOT)

Purely Public

Purely Private OperationMaintenance (OM)

Low

Design-BuildFinance-Operate (DBFO)

Private Sector Involvement

Build-OwnOperate (BOO)

High

Source: Kwak, Chih, and Ibbs (2009).

The CDC project departs from a conventional PPP in that the public sector is involved in all aspects of the project (design, build, finance, operate), but of‑ fers concessions (dark fiber) and service contracts (lit fiber bandwidth for use by ISPs, as well as maintenance contracts) to promote private sector participation. As will be discussed below in this chapter, under the terms of the dark fiber concession, the private sector is also required to invest in network expansion to reach new areas along the “access frontier” (second zone in Figure 13.4). In a parallel with these privately funded expansions, the state government continues to broaden the reach of the network by investing in areas with low population density and/or low incomes (third zone in Figure 13.4). When the CDC project began in 2007, the CDC planners believed that public investments and public-private partnerships could play an important role in reducing the “broadband gap” in relation to more developed regions. Based on household level data from Center of Studies on Information and Communication Technologies (Centro Regional de Estudos para o Desenvolvimento da Sociedade da Informação – Cetic.br), in 2010 18% of households in Brazil had fixed broadband connections of 256 Kbps or higher, and only 5% of households had connections of 2 Mbps or higher. Broadband penetration was lower for the Northeast region: when the CDC began operating, only 7% of all households in the region had fixed broadband connections of 256 Kbps or higher, and only 1% had connections of 2 Mbps or higher (CGI.BR, 2011, Table A6, p. 404). Since 2010, both Brazil and the Northeast region have posted significant gains in fixed broadband penetration, as illustrated in Figures 13.6a and 13.6b. Similar gains were observed in mobile Internet penetration.

318


Figure 13.6a. Brazil: Internet and Fixed broadband Penetration, by Household Percent of Private Homes with Service

50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 2006

2007

2008

2009

Internet penetration: Brazil

2010

2011

2012

2013

2014

Fixed broadband penetration: Brazil (>256kbps)

Fixed broadband penetration: Brazil (>2mbps)

Source: Developed based on data from Cetic.br (CGI.BR, 2007-2015)

Figure 13.6b. Northeast Region: Internet and Fixed Broadband Penetration, by Household Percent of Private Homes with Service

50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 2006

2007

2008

2009

Internet penetration: Northeast

2010

2011

2012

2013

2014

Fixed broadband penetration: Northeast (>256kbps)

Fixed broadband penetration: Northeast (>2mbps)

Source: Developed based on data from Cetic.br (CGI.BR, 2007-2015)


319

Despite these gains, recent data indicate that broadband in Brazil contin‑ ues to lag behind other countries in Latin America (see Chapter 1). Data from Cetic.br also indicate continued disparities within Brazil with regard to broadband access, usage, and connection speeds. These include ur‑ ban-rural gaps, regional differences, as well as major disparities across in‑ come groups. As indicated in Figures 13.6a and 13.6b, despite recent gains, the northeast region lags behind other regions of the country in Internet ac‑ cess and fixed broadband penetration. At the time when the CDC was being developed, there was mount‑ ing evidence on the impacts of investment in broadband access for eco‑ nomic development. In Ceará, broadband expansion is expected to lead to specific outcomes and long-term impacts including the rapid increase in Internet access for households and firms, more effective government administration, state budgetary savings from reduced OPEX payments to telco providers (leading to greater government capacity to invest), and improvement in public services across several critical areas (education, health, public security, and tax collection). These outcomes will lead to im‑ provements in quality of life and accumulation of human capital, as well as enhancements to the business environment (competitiveness, innova‑ tion capacity, and industrial diversification). Ultimately, these outcomes will lead to longer term economic impacts such as higher productivity, economic growth, and increased welfare. Figure 13.7 summarizes the “the‑ ory of change” underlying the CDC project:

320


Figure 13.7. CDC Theory of Change CDC Broadband Network

INPUTS

Better public policies and programs, with greater reach

Mass expansion of Internet use by the population

Expansion of education with better quality

OUTPUTS

Expansion of health More available and services with better effective public quality safety

Better standards of living and human capital

Improvement of Expansion of digital government resources available management and to companies investment capacity

R E S U L T S

Improvements in business environment: Competitiveness and innovation capacity

Economic diversification and increased productivity

I M P A C T S

Economic growth and increase in welfare

Source: BMR Consultoria e Treinamento (2012), p. 28.

It is worth mentioning that the CDC is the only statewide optical network in Ceará, and therefore the only source of fiber optic access for remote municipali‑ ties. In the past, the lack of quality data transport infrastructure was a major im‑ pediment for the implementation of the federal “Digital Cities” program in Ceará, a program from the Ministry of Communications focusing on improving public administration and digital services across Brazilian municipalities.4 Some of the results of the CDC initiative are already visible: these include the realized budgetary savings, growth of Internet use across the state, and expansion of public services that rely on information technology. Further, the broadband infrastructure has enhanced the innovation capacity of local IT firms in Fortaleza and elsewhere in the state. The government of Ceará has

4

See . Accessed: February 27th, 2016.


321

initiated impact evaluation studies to measure specific results and long term impacts of the CDC project.

PHASE III CONCESSION: THE CDC AS A PUBLIC-PRIVATE PARTNERSHIP During Phase III of the CDC development, Etice focused on a more struc‑ tured concession/PPP model to lease excess capacity to private sector oper‑ ators. This section and the next discuss the concession/PPP model, and the adaptations that were made to encourage participation in the auction for CDC dark fibers. The concept of a public-private partnership encompasses a wide variety of collaborations between government and private firms. The typical PPP specifies private ownership of infrastructure that is usually built by a private entity and then eventually transferred to the government agency that authorized its con‑ struction. At one extreme, a private entity undertakes all stages of infrastructure deployment by designing, building, financing and operating the facility. The CDC departs from this norm in a number of ways. To begin with, the government of Ceará designed, built, and financed the CDC fiber net‑ work. Those tasks, usually delegated to one or more private companies, were retained by the government. The International Monetary Fund (IMF) uses a broader definition of a PPP, treating private sector operation of a govern‑ ment-owned asset as a PPP.5 Accordingly, we refer to the CDC project as a public-private partnership because a collaboration between the government and private operators lies at its heart. The earliest examples of private-public partnerships – as they are known today – appeared in developed countries, and they concentrated on transpor‑ tation and utilities infrastructure projects. In time, the use of the arrangement spread to other markets and other services. Brazil established a federal PPP law in 2004 (Law n. 11,079), providing a legal and regulatory framework for PPP projects in the country.6 The first application of Brazil’s PPP law was the Metro Line 4 in the city of São Paulo (DELMON, 2010). This extension to the existing subway routes was built by 5

See IMF (2004). That arrangement has been classified as an “operating lease.” When the private entity is responsible for maintenance and improvement as well, which is the case with the CDC, it is described as a “concession” even if the entity does not necessarily sell services directly to the public. Also see PPPIRC (2015).

6 The PPP law in the State of Ceará, Law number 13,557, was established soon after in December 2004.

322


the state of São Paulo with funding from Inter-American Development Bank and several major international banks. As with the CDC, the design, build and finance phases were all undertaken by government agencies. A consortium of private firms operated the line drawing on transit fares as the principal source of revenue. Subsequent projects in Brazil adopted some form of public-private part‑ nership, including one notable ICT project – the data center built for use by the Banco do Brasil and Caixa Econômica Federal in Brazil’s capital, Brasilia.7 The CDC is the first communications network created in Brazil using a PPP model, though it is strictly speaking a concession rather than a PPP, under the specific federal and Ceará state laws. An arrangement that shares several of the CDC’s distinct features is the SuperNet in Alberta, Canada. That broadband network consists of over 13,000 kilometers of trenched fiber optic cables and 2,000 kilometers of high-speed wireless links. Bell Canada owns and operates a portion of the network that connects 27 cities in Alberta. The provincial government owns the remainder of the network that connects 402 communities in rural areas. The Alberta SuperNet is what sometimes is called “condominium dark fiber network” since the occupants of the fiber network hold the equivalent of “indefeasi‑ ble rights of use” (“IRU”) to a portion of the bandwidth (ST. ARNAUD and MACNEIL, 2001). While Etice built the CDC to fill the gaps in private investment in broad‑ band infrastructure and to reduce the cost of its own data transmission re‑ quirements, it recognized the powerful incentives of private partners to find cost efficiencies and to innovate on digital services. In particular, Etice took advantage of such incentives through use of an auction mechanism to engage private interests. The CDC’s procurement mechanism attempted to allocate the fiber capacity to providers of end-user services that are most efficient and most innovative relative to market alternatives. These providers would offer telecommunication services to the final users, including companies, house‑ holds, and the government itself. Ceará’s state government viewed the CDC as the only means to bring the digital age to underserved and unserved seg‑ ments of Ceará’s population.

7 See . Accessed: October 30th, 2014.


323

Basic Economics of an ICT PPP The various types of public-private partnerships constitute attempts to capture the benefits of the two forms of ownership while escaping their re‑ spective disadvantages. In this section we briefly identify characteristics of a PPP that affect its likely success or failure. The resulting pros and cons of PPP design are then used to evaluate the arrangement that was ultimately used in the case of the CDC.

Structuring a PPP: defining responsibilities and allocating risks Any infrastructure project requires the completion of a sequence of tasks (i.e., design, construction, financing, operation, maintenance and moderniza‑ tion), each of which could be carried out by the government or by a private entity. An important issue when organizing a PPP is whether to bundle those tasks and assign them to a single private firm, or to unbundle them so as to allow multiple contractors. As mentioned, Etice designed, built and financed the CDC, whereas private interests are contracted to operate, maintain and modernize the original fiber network. When contracting companies to complete the tasks, various criteria should be considered. Presumably the government’s goal in forming a PPP is pursuit of social welfare which subsumes cost and production efficiency. In so doing, however, it should acknowledge the market incentives that motivate potential private contractors. Indeed, the rationale for the CDC was a market failure that resulted in too little profit to justify a private investment to capture the net social benefits. The CDC had the secondary goal of promoting competition in the data transmission market under the assumption that competition would, first, capture greater benefits for the rural user population, and second, help cover Etice’s operations and maintenance costs for the CDC network. In ef‑ fect, as we describe in the final section of this chapter, the government itself benefited from this increased competition: at the end of 2015 the state govern‑ ment launched a tender for supplemental telecommunications services; the tender had ample participation from private sector bidders, and government costs were reduced by 60% as compared to the previous contract with a tele‑ com operator that previously held a monopoly position. Many considerations arise when deciding on how to structure the PPP arrangement. First, what are the relative capabilities of the public and private parties in terms of technical know-how, construction and operation exper‑ tise, and financial resources? Second, how able are the parties to cope with

324


the various sources of risk, including operational, competitive, political, and regulatory risks? Third, are there positive or negative externalities that arise between successive tasks that could be internalized by bundling the tasks? Fourth, how should the public and private partners share in reductions in op‑ erating costs that result from technological progress (IOSSA, MARTIMORT e MENESES, 2015)?

The role of government: risks and advantages Although Etice sought to correct for what it saw as a market failure, it recog‑ nized that governments are prone to failures as well. Arguably the most import‑ ant weakness of government participation in a PPP is that governments lack strong incentives when making economic and financial decisions. The govern‑ ing political party cannot help but respond to political and electoral pressures, and it is bound by administrative law and bureaucratic process. Authorization and funding may not extend beyond the next state or federal elections. Some institutions may severely limit the government’s range of action, such as when laws demand that equipment be domestically manufactured. At the same time, excessive discretion by the government partner can be another source of government failure. Weak administrative laws and minimal oversight can create a possibility that the public agent will be “captured” by private interests. Especially for projects that extend over many years, where the government must work closely and continuously with the private part‑ ners, there is a tendency to show favoritism that could be incompatible with the original social goals of the project. Governments nonetheless bring many unique advantages to a public-pri‑ vate partnership. Prominent among these is its pursuit of an equitable allo‑ cation that may not be realized by market outcomes. Another advantage is the means to curb abuse of market power of monopolies and oligopolies in the provision of telecommunications services, usually through competition laws and agencies. A further advantage is the ability to establish long-term plans and coordinate different actors (public, private, and academic) involved in a complex program such as the CDC. Governments can not only build an infrastructure such as the CDC, they can also use direct means to ensure that services are available to specific populations (e.g., rules on nondiscriminatory prices or subsidies), or they can work indirectly to increase competition in the downstream markets (e.g., lowering entry costs for firms or providing public infrastructure and telecommunications services to induce competition).


325

In addition, governments can use their fiscal powers to obtain financing at lower rates than private participants. Similarly, governments often can use their control of the legal system and regulatory policy to divert resources to the benefit of households and businesses located in these remote areas. The state’s legal power can be used to ensure contracting that supports investment by mitigating incentives for “hold up” and avoid the costly possibilities of renegotiation. Etice structured the CDC in a way that takes advantage of the relative strengths of public and private partners, but at the same time minimizes poten‑ tial conflicts of interest. Etice owns, operates, and maintains the core fiber ring of the CDC; it also owns last-mile facilities including local fiber and WiMAX fixed wireless for government use only. Concessionaires lease a portion of Etice’s network but they also own facilities; these include last mile distribution and also any fiber extensions added during their tenure. The ownership struc‑ ture also provides incentives for the private partners to innovate – finding new ways to reduce costs and improve efficiency – as they retain the benefits from such innovations.

PHASE III: CHALLENGES FOR THE IMPLEMENTATION OF THE CDC CONCESSION/PPP This section discusses the challenges and successes in implementing the CDC’s innovative PPP model. Our central focus will be on the develop‑ ment of the request for proposal (RFP) document for the concession (edital in Portuguese), which specified the bidding process and terms and conditions of participation, as well as the consultations carried out with potential private sector partners leading up to the auction. The edital stipulated the mechanics of the bidding process as well as the rights and responsibilities of winning bidders. We will show how adjustments made to the concessions document promoted greater participation by the private sector. This process took place without abandoning the dual objectives of ensuring the financial sustainabili‑ ty of the CDC and expanding broadband coverage in the state.

First Round of the Edital Prior to issuing the initial edital, Etice carried out a round of consulta‑ tions with various potential private sector partners, including large telecom operators and smaller Internet Service Providers. Etice issued the initial

326


concessions document in February 2013. That edital offered 15-year conces‑ sions for dark fiber in the CDC, with the option of renewing the contract for another 15 years. The offer was divided into three lots of fiber, including use of the optical ring, the linear extensions of the network, and the last mile links in cities and towns. The initial auction was conducted in April 2013. Despite the interest ex‑ pressed by potential private sector partners, no formal bids were submitted. Subsequently, both larger telcos and ISPs approached Etice with informal proposals to establish a partnership to use and maintain the CDC network.

USTDA Support and Consultation with Potential Private Sector Partners In 2013, the U.S. Trade and Development Agency (USTDA) provided funding for Etice to engage a team of consultants to assist in revising the first edital. The goal of the USTDA project was to refine the business model and concession document to attract private sector bidders and promote participa‑ tion of these companies in a new auction for CDC dark fibers. The U.S.-based company Developing Markets Group (DMG) was selected to collaborate with Etice, and formed a multidisciplinary team of economists, engineers, public policy specialists, and legal advisors to address the needs of the project.8 The project involved analyzing the terms and conditions of the initial edital, and discussing ways to restructure the terms based on international best practices, as well as considering the context in Ceará. An important step was to interview companies from the private sector in the state to understand their interests and to identify the factors that led them not to participate in the initial bidding. The team interviewed several companies active in the state ICT sector including small, medium and large operators, IT companies, and sectoral associations, among others. The interviews proved informative. First, they confirmed that mid-sized ISPs in the state had developed significant telecom capabilities, already served many markets in the interior of the state, and had a keen interest in leveraging the CDC network to improve and expand their services (including reselling bandwidth to smaller ISPs). Many of these ISP’s had been “trail-blazers” in the state, investing in their own networks, including fiber, and providing a range of services such as rural telecommunications, fiber to home, wireless data, point-multipoint radio service, and IP media distribution. Using the 8

Three of the authors (Feferman, Knight and Woroch) were members of this consulting team.


327

CDC would allow them to reduce their reliance on large operators that provid‑ ed access to many areas of the state. The ISPs understood the consumer markets in smaller towns and rural areas, and saw these markets as attractive business opportunities, with recent growth driven by rising incomes and the Bolsa Família federal cash transfer program.9 Hence, these ISPs were identified as strategic private sector part‑ ners that shared the same objective as the government of Ceará: expanding broadband connectivity in the interior of the state. Furthermore, even before the second edital was issued, these ISPs had begun collaborating more closely with one another, sometimes sharing their infrastructure. Second, the interviews confirmed the interest of the larger telecom oper‑ ators in using the CDC network. While these operators had built their own networks in the state, those networks did not provide sufficient coverage or redundancy in certain areas, particularly in the western and northeastern re‑ gions of the state. Operators also noted that the CDC could be used to pro‑ vide backhaul for mobile phone networks.10 One operator observed that it paid other telecom companies to utilize their broadband infrastructure, and would prefer avoid this dependency and expand its own network, leveraging the CDC to reduce costs. Larger operators reaffirmed that the interior of the state offered attractive business opportunities, with consumers increasingly using applications that require high bandwidth, such as video streaming and IP TV. In many areas, mobile broadband over 4G networks was becoming increasingly popular, but in the words of one operator, “at the moment fixed Internet arrives, consumers will abandon mobile Internet” as a substitute for a home Internet connection. (This is because the cost of data plans for mobile broadband remains very high in Brazil, as discussed in Chapter 1.) In sum, the value of the CDC network was evident across the interviews with potential private sector partners. Given this interest in the CDC, what were the main obstacles to the participation of the private sector in the fist auction? The interviews validated some of the hypotheses held by the Etice/ DMG team but also provided additional insights that offered guidance to revise the edital. The following themes and issues were emphasized in the

9

Bolsa Família is an important social welfare program in Brazil. It provides cash transfers to lower income families, conditional on the families maintaining their children in school and with their vaccinations up to date.

10 Chapter 2 in this book defines backhaul networks.

328


conversations with private sector companies, and are likely relevant for simi‑ lar public-private partnerships for broadband expansion in other regions: • Upfront payments: The size of the upfront payments to the state of Ceará in the initial edital was seen as a major impediment, by both larger operators and especially by smaller ISPs. The first edital had specified minimum upfront payments of about R$8 million to R$18 million, de‑ pending of the lot of optic fibers (this was equivalent to approximately US$4 million to US$8 million in mid 2014). Companies noted that they would prefer to use the funds to expand their own networks, or would prefer to carry out fiber swaps or smaller payments to other operators to expand the reach of their networks. Operators also mentioned that ad‑ ditional investments in equipment would be necessary to light the CDC dark fiber, further increasing the initial capital requirements. The size on the upfront payments in the initial edital discouraged participation, especially by smaller ISPs who were in many ways ideal partners. • Financing for upfront payments and other investments: Although financ‑ ing is available from public and private sources in Brazil, two issues were emphasized. First, smaller companies often have higher financ‑ ing costs, particularly as they find it harder to access low-cost loans from the Brazilian national development bank (Banco Nacional de Desenvolvimento Econômico e Social – BNDES). Second, many of the ISPs had already made significant investments in their networks (or would need to make future investments in equipment to light the CDC dark fiber), therefore these companies could not take on additional debt to fund the up-front payments, which would have required collat‑ eral. In the words of one ISP, “everybody is already fully invested.” • Network build-out requirement: The initial edital included a specific re‑ quirement for network build-out of 50 kilometers per year. This was seen as onerous by some prospective builders, who preferred to ex‑ pand their networks based on their own business planning and mar‑ ket opportunities (often, investments are carried out in cycles and can well exceed 50 kilometers per year). ISPs asked if they could obtain a “credit” for building more than 50 kilometers in any given year. The build-out requirement would also mean that Etice would need to supervise the private partners to ensure that the build-outs were undertaken, and there would be no assurance that these build-outs


329

would expand the network to areas that were high priority for the state government. • Revertibility and property rights: The initial edital stipulated that private investments to expand the CDC network be reverted to the state gov‑ erment of Ceará at the end of the concession period. This “revertibil‑ ity” of investments was seen as a major impediment by both smaller and larger operators, who wished to maintain the property rights over these network expansions. In addition, the revertibility requirement in the initial edital would not provide incentives for private partners to build the best possible network since they must eventually transfer the additions to the state of Ceará. • Network maintenance and operation: Some potential partners were crit‑ ical of provisions for network maintenance in the initial edital that called for lessees to share the responsibility of maintaining the CDC. This shared approach could lead to coordination challenges, potential “free-rider” problems, and unforeseen costs and risks. One operator suggested the use of a single maintenance contract; ultimately, Etice adopted this approach when the edital was revised, utilizing a separate edital for the network maintenance. Other companies also mentioned the perceived high cost of the monthly network management fee (R$205,000 per month, or about US$92,000 in mid 2014) included in the initial edital. • Provision of basic Internet access: The initial edital included the require‑ ment to provide low-cost basic Internet access at R$29 per month (about US$13 per month in mid 2014) for a 1 Mbps connection, as specified in the National Broadband Program (Programa Nacional de Banda Larga – PNBL). Operators had mixed reactions to this require‑ ment. One operator noted that this requirement would reduce flexibil‑ ity to run his business. Other operators noted that they already offered the basic plan, but that most customers prefer more expensive options with higher connection speeds and greater data allowances. Yet anoth‑ er ISP emphasized that many of his customers could only afford the low-cost plan, which it would continue to offer. • Political risk: Some of the companies interviewed mentioned the risks that arise with changes in government. One company noted that, de‑ spite the guarantee of a 15-year contract (with a 15 year renewal option),

330


a new government with different priorities could neglect certain parts of the network. Thus, while the CDC public-private business model was structured to ensure its financial sustainability (providing a mea‑ sure of insulation from political vagaries), some potential partners re‑ mained cognizant of risks. In summary, the interviews indicated that the CDC would provide signif‑ icant value to private sector partners in Ceará, who were keen to expand their services in the state. But the high up-front costs and burdensome require‑ ments of the initial edital (annual network expansion of 50 km, the revertibil‑ ity of investments, and shared maintenance) had deterred the participation of these potential partners.

The Revised Edital Based on the analysis and consultations with the private sector, the Etice/ DMG team reformulated the concession document. The revised edital reduced the burden on successful concessionaires with the goal of stimulating interest in participation, but without compromising Etice’s fundamental objectives. Changes included transforming upfront fees into monthly payments (to reduce financing requirements), elimination of the network management fee, sharing of network maintenance costs (outsourced under a separate contract), relaxing of the build-out requirements and revertibility terms, removal of the universal service obligation (offering the basic plan under the PNBL called Popular Broadband (Banda Larga Popular)), and expanded access to Coelce rights of way (through the existing agreement between Etice and Coelce). To promote competition, more fiber pairs (10 instead of 8) and more lots (4 instead of 3) were assigned for concession, and no company or consortium could win more than one lot. Some of the lots were structured to be attractive to smaller ISPs, with a smaller number of fiber pairs on the main ring and last miles deployments in fewer locations. One important principle followed by the revised edital was to rely more heavily on market-based incentives: the new edital allowed operators more flexibility to plan their own investments based on market opportunities in the interior of the state (where they were already investing) and removed the revertibility requirement to ensure that operators could retain property rights for privately financed network expansions. Etice would carry out subsequent public investments in areas that remained under-served, and would gain ac‑ cess to 2 pairs of fibers in all privately built network extensions. The revised


331

edital continued to promote widespread access to broadband, while fostering broadband service competition. As mentioned above, ten dark fibers were put up for bid in four lots of fiber pairs (one lot with two pairs, and three lots with one pair). The lot winners were to be the companies or consortia offering the highest monthly rental value. No concessionaire was allowed to win more than one lot. Telecoms carriers could propose joint bids with any kind of company. While Etice would be responsi‑ ble for maintaining the network (through a separate contract), concessionaires would light the fibers with their own equipment and maintain this equipment. The winners would make monthly payments to cover OPEX of the network, guaranteeing its long-term financial sustainability under this PPP. The restructured edital significantly reduced the requirements for private sector operators, providing more incentives for participation. At the same time, the terms of reference supported Etice’s goals of achieving financial sustainability for the CDC and promoting digital inclusion for the population of Ceará. The expansion of broadband coverage would happen in four ways: • Private investments for network expansion based on business opportunities; • A modest requirement for operators to offer access to 10 new towns and cities, to be chosen by the operators themselves from a list of 84 towns and cities; • Etice’s rights to 2 pairs of fiber in network expansions carried out by private partners, which would extend the government network; and • Subsequent public investments to expand the network to areas that remained uncovered. As Etice eventually learned, a strategy of “build it, and they will come” may not necessarily work. The ultimate success of a tender of infrastructure capacity, as the one undertaken by Etice, depends critically on whether credi‑ ble bidders find it in their own interest to participate. Many factors affect the decision of potential bidders to participate in the auction and rational bidders make a complete life-of-project assessment of the benefits and costs, includ‑ ing the bidding costs (technical evaluation, financial analysis, and proposal preparation), initial investments, and the implementation of the project over 15-30 years.

332


Results and Insights from the Second Auction Etice incorporated many of the changes recommended by the DMG team, which focused on the specific requirements that had discouraged participa‑ tion by private sector partners in the 2013 auction. The state of Ceará judicial branch approved the revised auction documents during the fourth quarter of 2014, and the new round of concession bidding took place on December 8, 2014, in Fortaleza. Three of the main local ISPs operating mainly in the state of Ceará partic‑ ipated in the auction as a consortium: Brisanet, Mob Telecom, and Wirelink. They submitted a joint bid on Lot Two for R$150.000 in monthly lease pay‑ ments (equivalent to US$ 58,000 at the end of 2014). While participation was lower than expected, with only one consortium bidding, the outcome was generally viewed as positive. First, the three companies involved in the bid represent ideal partners for the state of Ceará, given their expertise the regional markets and their demonstrated willingness to expand services to the interior of the state. The firms are based in Ceará, pay taxes in the state, utilize well-qualified technical staff, and hire their employees locally. Second, the participation of three com‑ panies supports the goal of promoting market entry and competition: while the three companies in the consortium will share the CDC infrastructure, they will compete in target markets amongst themselves, as well as with other firms. The three companies will also likely sell capacity to downstream ISPs, further promoting market entry and competition in local services provision. Finally, the monthly lease payments from the concession will be sufficient to cover the maintenance cost of the entire fiber network, supporting the finan‑ cial sustainability objective of the CDC. The outcome of the second auction highlighted two aspects of the bid‑ ding process. First, Etice pursued an indirect approach to “set-asides” when it structured the fiber lots to differentially appeal to smaller ISPs and larger tel‑ cos. In effect, the consortium of smaller ISPs found the lower cost Lot Two – with just one pair of fibers in the main ring – the most attractive option given its needs. The consortium may not have participated if the lots had been struc‑ tured with more fiber pairs and with a correspondingly higher minimum bid. Second, Etice expressly encouraged potential bidders to form consortia to bid at the tender. This strategy was adopted cognizant of the potential to reduce competition, but Etice felt the risk was acceptable to promote the participation of regional ISPs that would not have the scale to proceed alone.


333

We also note that the larger telecom operators did not to participate in either of the two auctions. Interviews with some of these operators prior to the second auction indicated several possible reasons. Some operators had already undertaken their own investments in network expansion in the state, and consequently the CDC concession may have become less attractive over time, even at a reduced cost. Corporate financial reporting was another con‑ sideration: two potential bidders indicated that operating leases (such as in the case as a concession with monthly payments) could be problematic because they would reduce reported net income. These operators mentioned that they would prefer to engage in infrastructure swaps, which have a lower impact on cash flow and on corporate earnings. Additionally, some telcos were going through a process of corporate restructuring or merger in late 2014, which impacted new initiatives. It was also clear from the interviews that new in‑ frastructure commitments would require a lengthy internal review process within these companies, with decisions made at central headquarters outside of Fortaleza and involving multiple stakeholders. Finally, other markets and projects in Brazil may have had higher priority at the time. Nonetheless, the leadership of the consortium of regional ISPs that took part in the auction indicated that the three participating companies had been contacted by mobile phone operators who were interested in expanding their 4G networks using the CDC fibers and data transport equipment for their backhaul. Therefore, some of the larger telcos may in effect leverage the CDC infrastructure, even if they did not participate directly in the auction. The con‑ sortium estimated that it would invest R$15 million (US$6 million based on the exchange rate at the end of 2014) in capital expenditures to activate the fibers and another R$20 million (US$8 million) in last mile investments, in‑ cluding fiber-to-the-home GPON technology for cities with more than 20,000 inhabitants. In the future, Etice may offer new concessions for the remaining pairs of available dark fiber in the CDC network, which it continues to expand in underserved areas of Ceará. With the expected growth in bandwidth demand, it is entirely possible that Brisanet, Mob Telecom, and Wirelink would bid for new lots, either as a consortium or individually. Furthermore, other potential concessionaires will have more visibility into the experience of these three “first movers”, and may decide to enter the wholesale market through future CDC fiber concessions. These new entrants will need to weigh the benefits and costs of investing in dark fiber, in relation to buying existing capacity

334


from large telecoms or from the three current CDC concessionaires. This cal‑ culus will depend on the pricing for this existing capacity as well as on the cost of new CDC dark fiber concessions. As noted earlier in this chapter, in the past Etice had depended on a mo‑ nopolistic regional operator (the company Oi) for government telecommuni‑ cation services. The development of the CDC replaced most of these services (at a significantly lower cost), but the state government continued to rely on Oi for supplemental telecom services, particularly in areas without CDC cov‑ erage. In late 2015, a new tender was issued for these supplemental telecom services. This time, eight qualified companies participated in the bid, and the contract was awarded in January 2016 to two of the companies that had leased the CDC excess capacity. The broad participation in the tender and the amount of the winning bid (60% below the previous contract with Oi) indi‑ cate the extent to which the CDC concession has promoted competition for telecommunications services in the state. The successful completion of the PPP concession process in Ceará, and subsequent developments, are important steps in validating the CDC busi‑ ness model for other states in Brazil, which continue to pursue similar initia‑ tives. At the end of 2015, other state IT companies continued to reach out to the Etice team for guidance in deploying their own broadband networks.

LESSONS LEARNED Here, we summarize the various lessons we extract from the experience of Etice’s public-private business model for implementing the CDC.

1. Leveraging partnerships and the strengths of each partner The CDC business model bridges the gap between purely public and purely private ownership, leveraging the strengths of partners in the public, private, and academic sectors. The initial CDC deployment benefited from its collab‑ oration with the state electrical utility that provided access to its rights-of-way and utility poles. The academic community was also a key partner during the CDC’s “proof of concept” phase. In the current phase, private sector telecoms and ISPs bring a range of capabilities to the public-private model including rapid deployment of digital services, marketing reach, innovation, and attrac‑ tive pricing. Based on business incentives, and leveraging the government


335

built CDC infrastructure, these private sector partners are expanding service to the “access frontier”, described earlier. Etice’s core business is managing the fiber backbone infrastructure while delegating data transport services, network maintenance and last-mile dis‑ tribution to private sector partners. By including many private sector partici‑ pants, Etice pursues its goals of expanding network coverage and improving redundancy, while at the same time promoting low prices and innovative ser‑ vices for the population of Ceará. The state government made three important contributions to the CDC: (i) the initial investment in the network infrastructure, (ii) the management of the shared backbone infrastructure, and (iii) the coordination of the different actors in the ecosystem, which includes several private sector, academic, gov‑ ernment partners. The public investment in the infrastructure has several benefits. It takes advantage of the public sector’s lower cost of capital, relieving all carriers (especially the smaller ISPs) of the CAPEX burden. It allows Etice to use a simpler auctions/concession PPP, rather than the more complex Brazilian PPP Laws (federal law n. 11.079/2004 and state law n. 13.557/2004). The ar‑ rangement under these laws is tailored to the usual “build-operate-transfer” scenario and requires privately financed investments and revertibility of new investments at the end of the project – a major obstacle for participation of private sector operators in the initial concession bidding.

2. Balancing financial sustainability and social goals The CDC balances the requirement for financial sustainability with social goals of providing broadband access to unserved and underserved inhabitants of the interior as well as in major cities in Ceará. With regard to sustainability, the CDC model is structured to generate the revenues required for operations, avoiding financial dependence, which buffers the CDC from the changing winds of state politics and changing budgetary allocations. As noted above, the CDC has lowered the state government’s expenses significantly by dis‑ placing connectivity contracts with the incumbent private sector telco (Oi), or by renewing contracts at significantly better terms with new operators in the market. Etice also charges government clients for its services according to a transparent fee schedule. Nevertheless, the state of Ceará struggles with the temptation to provide free broadband services to its government clients, a practice pursued by many other Brazilian states. The proposition of “free

336


Internet for everyone, everywhere, all the time” spells financial doom for a project such as this. The dark fiber concession is another element of the financial sustainabili‑ ty strategy. Through the concession, private operators contribute significantly to the long-term financial sustainability of the CDC by sharing operational costs. Lower OPEX for the government allows more investment in the long term, including new investments in the expansion of the CDC. At the same time, the CDC stays focused on its mission to promote universal broadband access in the state of Ceará. This objective was incorporated in the second edital, which leased excess dark fibers in the CDC network to private sector concessionaires, but with requirements and incentives for the private expan‑ sion of the network. Indeed, the CDC has already made important contributions to digital in‑ clusion in the state. Etice’s main metric used for its planning decisions – the “number of users connected” – shows an increase over time. For example, Etice tracks the number of email addresses issued in connection with the CDC, which reached nearly 1.5 million in mid-2014, representing 17% of the estimated population of Ceará. A World Bank analysis, reported by Etice in 2014, indicated a broadband availability rate for urban population in the state of 82% (this coverage was virtually nil prior to the CDC).

3. Promoting competition Competition in the telecoms sector is critical to keeping rates for broad‑ band affordable. Lower prices are the best way to increase Internet adoption and use. In a competitive market, scaling up usage drives revenue, and ul‑ timately supports greater investment. The CDC model can help stimulate competition in two ways. First this model offers an alternative network to the private sector incumbent carriers, which may be operating as a monopoly or oligopoly. Second, the CDC model can promote competition by facilitating the entry and expansion of smaller carriers into the market. In Ceará, the smaller ISPs have been instrumental in reaching customers in the interior of the state. In contrast, large telcos could use the CDC infrastructure as a means to deliver service other Brazilian states (leveraging Fortaleza’s strategic position as a landing point for international cables), rather than expanding Internet access in underserved regions of Ceará.


337

EXPORTING THE CDC MODEL Above we described the ongoing efforts and challenges in Brazil to reform its telecommunications sector. We also noted the urgent need to expand and modernize the country’s ICT infrastructure so as to increase productivity, im‑ prove the delivery of public services, and promote digital inclusion. Both the private sector and public/academic institutions have been involved in broadband expansion in Brazil, but with limited collaboration between them. On the private sector side, large telcos have expanded access at scale, especial‑ ly in more profitable urban areas, and met some requirements for providing broadband coverage to public schools Nevertheless, problems persist with re‑ gard to broadband coverage, service quality, and cost of service. The govern‑ ment and academic broadband networks currently deployed (most of these or‑ ganized by the RNP) do not follow a commercial model. A broader utilization of PPP arrangements such as the CDC, involving these government and academ‑ ic networks, could scale coverage well beyond their academic and government users. As seen previously, this strategy could also promote competition in the telecoms sector. Where local public/academic networks are not present, gov‑ ernments usually rely on expensive contracts with private carriers for their own connectivity requirements. The CDC case illustrates how the public, private and academic sectors can collaborate to expand coverage in the “access frontier”. Many Brazilian states face the same initial challenges as Ceará: high-cost of broadband service and high concentration among regional telecom carri‑ ers. Several Brazilian states are searching for models of public-private collab‑ oration to reach underserved regions in a country of continental proportions. Because the laws that regulate telecommunications are federal, the CDC mod‑ el can be applied to other states. States such as Rio de Janeiro, Pará, Paraiba, Bahia, Pernambuco, Rio Grande do Sul, Piauí, Sergipe, Tocantins, Santa Catarina, and Rio Grande do Sul are exploring PPPs inspired, in part, by the CDC experience. The goals of these projects are similar: expand broadband access to underserved areas, connect government agencies to improve the quality of public services, and achieve financial sustainability without relying on state budgetary allocations. During 2014, Etice discussed the CDC experience with officials from several Brazilian states, including Bahia, Minas Gerais, Paraíba, and Rio Grande do Sul, among other states. And during the 2014 FIFA World Cup, Telebras used the CDC infrastructure to fulfill FIFA’s connectivity requirements, further promoting the CDC nationally.

338


Conditions Conducive for the CDC Model The CDC model is most relevant for regions with significant unmet bandwidth demand, urban-rural service gaps, limited competition among carriers, and high prices for Internet access. In these regions, public in‑ vestments and business models along the lines of the CDC can stimulate competition, lower service costs, and promote access, while covering the state IT company’s operational costs. Studies by the Brazilian Ministry of Telecommunications indicate that reductions in service costs have a significant impact on telecommunications demand. The price elasticity of demand for telecommunications services in Brazil ranges from 1.0 to 3.4, and therefore a 10% reduction in service costs can raise demand by 10% to 34% (KNIGHT, 2014, p. 84-85). The presence of a vibrant community of ISPs can accelerate provision of broadband services. In regions where financing for small and medium enter‑ prises is costly, public investment in the infrastructure becomes especially im‑ portant. By financing the initial investment, the government can greatly reduce the CAPEX requirements of private sector partners that participate in the proj‑ ect under a simpler concessions structure (rather than a PPP under Law n. 11,079, as noted above). Additionally, the infrastructure can be shared by mul‑ tiple smaller carriers, as in the case of the CDC. Finally, partnerships with state electrical utility companies can greatly reduce the network deployment costs by leveraging existing infrastructure including poles, ducts, and substations.

Impediments to Implementing the CDC Model Where the political influence of the major incumbent telecommunications operators is strong, there can be resistance to the development of state-owned broadband networks, since this tends to reduce the monopoly or oligopoly rents accruing to these firms. National constitutions/laws may also restrict the formation of PPPs. Another impediment to exporting the CDC model is the implementation capacity of the public sector, including the financial, technical, and managerial resources required to undertake a large, complex telecommunications project. For example, at the beginning of 2016 many Brazilian states struggled with budgetary constraints, which made financing a project like the CDC more difficult.11 Finally, implementing a project such 11 It should be noted that lending to public sector projects by the World Bank, the InterAmerica Development Bank, and the Brazilian National Economic and Social Development


339

as the CDC requires the capability to coordinate the various actors (public, private, and academic) in the “ecosystem”, as demonstrated by the CDC case.

CONCLUSIONS The state of Ceará has made rapid progress in meeting its broadband needs by developing an innovative PPP business model to deploy the CDC network. The CDC is considered to be state-of-the art network, similar in many respects to the best examples of such networks in the United States. The CDC experience also shows that government financing and ownership of broadband infrastructure can facilitate the participation of private companies and, paradoxically, promote private sector competition. Further, a financially sustainable business model can insulate the project from political risks, such as changes in government – a challenge in both Brazil and in many develop‑ ing countries. Over time, Etice has adapted its model to respond to market conditions and the concerns of potential participants in its auction of fiber capacity. Other Brazilian states are closely studying the experience of Ceará as they plan to build and expand their own broadband networks. The CDC’s fiber optics infrastructure allows Ceará’s public administration to roll out innovative broadband applications, otherwise impossible, such as telehealth, distance learning, video surveillance for public safety, and remote cargo scanning on roads. Further, the CDC promotes greater efficiency in the state administration by expanding access to information technology to all state and municipal agencies. And finally, access to high-speed connectivity pro‑ motes business innovation in cities such as Fortaleza – a growing ICT cluster. The state government is collecting randomized baseline data to determine the impact of broadband access on schools, hospitals, public safety agencies, and private sector enterprises, among others. Analysis of this data should provide quantification of the benefits offered by increased broadband access. The CDC model serves as a guiding framework for municipal, state and national governments in Brazil and in other countries that face the challenge of sustainably expanding broadband access and improving the quality of pub‑ lic services. Bank (BNDES), is also constrained by state borrowing limits set under the federal Fiscal Responsibility Law (Lei de Responsabilidade Fiscal – Lei Complemetar 101/2000). See Accessed: October 30th, 2016.

340


References BMR CONSULTORIA E TREINAMENTO. Desenvolvimento do Sistema de Avaliação de Resultados do Programa Cinturão Digital do Ceará, Produto 2. Fortaleza, June 11, 2012. COMITÊ GESTOR DA INTERNET NO BRASIL (CGI.BR). ICT Households and Enterprises – Survey on the Use of Information and Communication Technologies in Brazil, volumes for the years 2005-2014. São Paulo: NIC.br and Cetic.br, 20062015. Available on: . Accessed: Februray 27th, 2016. DELMON, Jeffrey. Understanding Options for Public-Private Partnerships in Infrastructure, Sorting out the Forest from the Trees: BOT, DBFO, DCMF, concession, lease… The World Bank Finance Economics & Urban Department, Policy Research Working Paper #5173, January 2010. GOMES, Cid F.; GOMES, Fernando C.; FERNANDEZ, Marcial P. Infrastructure and Business Model for Universal Broadband Access in Developing Regions: The Ceará State Digital Belt. Third IFIP TC 6 International Conference, WCITD 2010, Brisbane, Australia, September 20-23, 2010. Available on: . Accessed: Februray 27th, 2016. INTERNATIONAL MONETARY FUND – IMF Public-Private Partnerships. Washington, DC: Fiscal Affairs Department, March 12, 2004. Available on: . Accessed: Februray 27th, 2016. IOSSA, Elisabetta; MARTIMORT, David M.; MENESES, Flavio. The Simple MicroEconomics of Public-Private Partnerships. Journal of Public Economic Theory, 17:1, 2015, pp. 4-48. Klemperer, Paul (2002). What Really Matters in Auction Design. Journal of Economic Perspectives: 16, p. 169–189. KNIGHT, Peter T. The Internet in Brazil: Origins, Strategy, Development, and Governance. Bloomington, Indiana: AuthorHouse, 2014. KWAK, Young Hoon; CHIH, YingYi; IBBS, C. William. Towards a Comprehensive Understanding of Public Private Partnerships for Infrastructure Development. California Management Review, 51: 2, 2009, pp. 51-78. PPPIRC. What are Public Private Partnerships? (2015). Available on: . Accessed: September 20th, 2016.


341

QUEIROZ, Cesar; ASTESIANO, Gaston; SEREBRISKY, Tomas. An Overview of the Brazilian PPP Experience from a Stakeholders’ Viewpoint. Inter-American Development Bank Infrastructure and Environment Sector Technical Note No. IDB-TN-64, March 2014. Washington, DC: Inter-American Development Bank. SAENGUDOMLERT, P.; MODIANO, E; GALLAGER, R. G. Gallager. On-line routing and wavelength assignment for dynamic traffic in WDM ring and torus networks. IEEE/ACM Transactions on Networking, 14: 2, 2006, pp. 330-340. ST. ARNAUD, Bill; MACNEIL, David G. Community Condo Fiber Networks: The Customer Empowered Networking Revolution. Canadian Advanced Research and Innovation Network (CANARIE), EDUCAUSE Policy Office, January 2001. WALLSTEN, Scott. Reverse Auctions and Universal Telecommunications Service: Lessons from Global Experience. Federal Communications Law Journal, 61: 2, 2008, pp. 373-394. Available on: . Accessed: Februray 27th, 2016.. WORLD BANK. Telecommunications Regulation Handbook, 10th Edition. Washington, DC: World Bank, 2011. Available on: . Accessed: Februray 27th, 2016.

| 14 |

UNDERWATER CABLES IN BRAZIL AND THE WORLD Eduardo Grizendi [ Brazilian Education and Research Network – RNP ]

Michael Stanton [ Brazilian Education and Research Network – RNP ]

Decílio Sales [ Brazilian Army ]

T

INTRODUCTION

his chapter discusses the essential issue of how telecommunications connexions between places that are geographically separated by bod‑ ies of water were and still are made. This became an urgent issue in the 19th century when electrical wires began to be used for communication and this urgency applies nowadays to their successors, optical fibers. In both cases, the solution is to place or bury at the bottom of the bodies of water the electrical or fiber-optic cables, usually shielded by external armor. The first application of underwater cables was carried out to interconnect European countries that were separated by the sea, but the objective of expanding the in‑ terconnected world soon led to more ambitious projects to interconnect con‑ tinents by submarine cables. From crossing the sea it was a small step to also use the same technology for laying cables at the bottom of lakes and rivers. In this latter case, they are known as subfluvial cables. We kick off this chapter with a brief description of the underwater electrical cables used in the 19th century to build the first global telecommunications network, which has been called the “Victorian Internet”. Today’s Internet depends intimately on a huge mesh of submarine fiber-optic cables to offer global coverage. As for Brazil, the text briefly describes the Brazilian participation in the Victorian Internet, with submarine electrical cables that were used along its

344

e. grizendi • m. stanton • d. sales

entire coastline, and highlights the subfluvial cable that extended along the Amazon River up to the city of Manaus. The second part of this chapter de‑ scribes the evolution of how submarine optical cables were used in Brazil from 1994 onwards and the projects of new cables that were underway in mid-2016. In the third part the text we describe a large project for using subfluvial optical cables to provide Internet access to the interior of the Amazon region.

A BRIEF HISTORY OF THE EVOLUTION OF UNDERWATER CABLES IN BRAZIL AND THE WORLD In the 19th century, the study of electricity led to the invention of electrical telegraphy1 and telephony2 that made instant communication feasible over long distances. In both cases, telegraphy and telephony started off with terres‑ trial communication, using metal electricity-conducting wires, usually hung from poles, to interconnect communication endpoints. Demonstrating the advantages to governments and businesses of (almost) instantaneous com‑ munication led to demands to extend it overseas, enabling communication with islands and other countries and continents. The solution for meeting these demands was invention of the underwater cable that encapsulated the wires, insulating them from contact with water, and also providing protection against mechanical damage by animals or through accidents. By the end of the 20th Century, the metal wires used in telecommunica‑ tions were replaced with optical fibers, even though the former are still pres‑ ent in long-distance underwater cables, to provide electricity to energize the optical signal amplifiers installed along the cable. In this chapter we will present the usage of underwater cables in the con‑ struction of the so-called Victorian Internet (of the 19th century) (STANDAGE, 1998), and their current role in worldwide networks and in the Internet in 21st century Brazil.

1

The first telegraphy patent was registered by Cooke and Wheatstone in England in 1837. Independently, Morse patented another mechanism in the US in 1843. Source: . Accessed: Nov 2nd, 2015.

2

Practical telephony is credited to Bell, who registered a patent in 1876 in the US. Available on: . Accessed: Apr 25th, 2016. The first transatlantic telephony service, using radio, was installed in 1927 and the first one to use a submarine cable was only deployed in 1956. Source: . Accessed: Apr. 25th, 2016.

underwater cables in brazil and the world

345

THE FIRST UNDERWATER CABLES FOR TELEGRAPHY IN BRAZIL AND IN THE WORLD The first experiments immersing metal wire cables in water were con‑ ducted in Europe from 1839 on.3 One of the main breakthroughs resulting from these experiments was the electrical insulation of the wires, using as insulating covering a glutinous substance extracted from a tree from Malaysia called gutta-percha. Starting in 1851 the first underwater cables for commer‑ cial use were laid, starting with a cable extended across the English Channel, a narrow stretch of the Atlantic Ocean that separates the island of Great Britain from the north of France (BRIGHT, 1898).4 Throughout the 1850s, many cables were laid in Europe, and between Europe and Africa, with distances that ranged from dozens to a few hundred kilometers. For commercial and strategic reasons, the great interest of the British was communication with their empire, spread through all the conti‑ nents. Thus the first transoceanic cable was launched in 1858 and connected Ireland to North America (Figure 14.1), a distance of approximately 4,000 km. Unfortunately, due to problems in the electrical design and to damage to the cable, the new connection stopped working after 23 days. A new attempt to lay a cable of improved design was made in 1865, and this also failed when the cable broke and could not be recovered from the ocean bed at a depth of 4,000 meters. Finally, in the following year the same company successfully lay a third cable that remained operational for 100 years.5

3 Submarine communications cable. Available on: . Accessed: Oct. 4th, 2015. 4 History of the Atlantic Cable & Undersea Communications from the First Submarine Cable of 1850 to the Worldwide Fiber Optic Network. Available on: . Accessed: Oct.4th, 2015. 5

A detailed description of the history of these three attempts is available on: . Accessed: Oct. 4th, 2015. See also the section “Thompson and Whitehouse” in the article “Mother Earth, Mother Board”, of Neal Stephenson, published in Wired magazine on Jan 12th, 1996. Available on: . Accessed: Oct 31st, 2015.

346


FIGURE 14.1. Map of the route of the 1858 transatlantic cable

Source: Frank Leslie’s Illustrated Newspaper (1858).

The first entirely submarine telegraph cables were laid in Brazil in the 1870s and connected the cities of Rio de Janeiro, Salvador, Recife and Belém. After these, in 1875, connections were made between the cities of Recife, João Pessoa and Natal (COSTA DA SILVA, 2011).6 At the same time, the first international submarine cables were laid, one of them connecting Recife to Carcavelos, in Portugal, passing by the islands of Cape Verde and Madeira, and the other one connecting Rio de Janeiro to Montevideo, in Uruguay. In 1893 the first submarine cable landed on the Brazilian island of Fernando de Noronha and, between 1895 and 1896, a subfluvial cable was laid along the course of the Amazon River, connecting Belem to Manaus (Figure 14.2). By the turn of the 20th century, the Victorian Internet had already been built, connecting a considerable part of the world by telegraph cables (Figure 14.3). The world had already become a lot smaller.

6

Costa da Silva (2011) describes and discusses the evolution of telegraphy in Brazil during the Second Empire (1840-1889), including the extensive usage of submarine cables.


347

FIGURE 14.2. Subfluvial telegraph cable between Belém and Manaus, laid between 1895 and 1896

Source: RNP, based on Siemens (1896).

FIGURE 14.3. Victorian Internet – Global telegraphy in 1901

Source: Eastern Telegraph

348


SUBMARINE OPTICAL CABLES IN BRAZIL As already mentioned, telephony only started using transoceanic subma‑ rine cables in 1956 when they became more attractive than radio transmis‑ sion. These cables initially used coaxial conductors with amplifiers. The first of these, called TAT-1 (Transatlantic No. 1), initially carried 36 voice channels and this number was later increased to 51. TAT-7, laid in 1978, carried be‑ tween 4,000 and 10,500 circuits. Nevertheless, starting with TAT-8 (1988), transoceanic cables adopted optical transmission, substantially increasing the capacity of the link to 40,000 circuits in this first case. It should be remem‑ bered that the Internet was launched in 1983 and its first commercial use was in 1989, and that Internet content transmission was – and still is – in capacity terms much more demanding than that of telephony. Hence data transmission demand started to determine capacity planning of the national and international communications networks. To understand the significance of TAT-8, we should observe that its 40,000-voice circuit of 64,000 bit/sec‑ ond capacity corresponded to a total of 2.56 Gbps (gigabits per second), a great breakthrough at the time, but quite small when compared with current tech‑ nology. Since 1988 optical fiber transmission technology has evolved a lot. For instance, in 2015 one single optical fiber is capable of transmitting in a single direction more than 100 information channels, each of which transmitting 100 Gbps, amounting to 10 Tbps (terabits per second), that is, 4,000 times the capacity of the TAT-8 cable. Typically, a modern underwater cable would have three or four pairs of fibers (each pair transmits in both directions). This evolution in optical transmission capacity not only made possible the explosive growth of the Internet in recent years, but was also its consequence. We can follow this evolution by observing Internet traffic growth between Latin America and the rest of the world in the same period, and also by the growth of the capacity of the cables used to transport this traffic. Below we present in a concise manner an overview of the usage of underwater optical cables by Brazil for national and, mainly, international communication. The data are presented in Table 14.1 and commented on in the text. Figures 14.4 and 14.5 show maps of most of these cables.


349

FIGURE 14.4. Map of existing and planned international optical cables involving Brazil in 2016

Source: Submarine Cable Map.

350


TABLE 14.1. Submarine optical cables, in use and planned, involving Brazil Length (km)

Landing points in Brazil

Other countries/ territitories served

0.56 Gbps in Brazil

8,116

Fortaleza

Venezuela, Trinidad and Tobago, US Virgin Islands, US

1994

0.56 Gbps

1,720

Florianópolis

Uruguay, Argentina

Embratel (Brazil)

1996

initially 0.622 Gbps

2,552

Santos and capitai cities between Rio de Janeiro and Natal

Americas-2

Embratel (Brazil), AT&T, Verizon, Sprint, CANTV, Tata Communications, Level 3, Centennial of Puerto Rico, Corporacion Nacional de Telecommunicaciones, Telecom Argentina, Orange, Portugal Telecom, C&W Networks, Telecom Italia Sparkle, Entel Chile

2000

7.5 Gbps

8,373

Fortaleza

Venezuela, French Guiana, US, Martinique, Porto Rico, Trinidad and Tobago, US Virgin Islands, Curaçao

Atlantis-2

Embratel (Brazil), Deutsche Telekom, Telecom Italia Sparkle, Telecom Argentina, Telefonica (Spain), Portugal Telecom, Orange, Telefónica Larga Distancia de Puerto Rico, AT&T, Belgacom, KT, SingTel, Sprint, Tata Communications, Verizon, BT

2000

40 Gbps

8,500

Fortaleza, Rio de Janeiro (use only by Embratel)

Argentina, Cape Verde, Senegal, Canary Isles (Spain), Portugal

Cable

Present owner(s)

Year

Capacity

Americas-1

All America Cables & Radio (Dominican Rep.), Antel (El Salvador), Antel (Uruguay), Antelco (Paraguay), Austria PTT, AT&T, Belgacom, BET (Barbados), BT, C&W, CANTV, Codetel (Dominican Rep.), CPRM, Dacom, DBP Telekom, Embratel (Brazil), Entel (Chile), France Telecom, GT&T

1994

Unisur

Telefonica (Spain), Antel (Uruguay)

Brazilian Festoon


351

(Continuação) Landing points in Brazil

Other countries/ territitories served

25,000

Rio de Janeiro, Salvador, Santos

Colombia, Puerto Rico, US, Guatemala, Ecuador, Peru, Chile, Argentina

3.84 Tbps

20,000

Fortaleza, Rio de Janeiro, Santos

US Virgin Islands, Venezuela, Colombia (Level-3 only), Panama, Peru, Chile, Argentina

2014

50 Tbps

17,800

Fortaleza, Salvador, Rio de Janeiro

Dominican Rep., Puerto Rico, US, Mexico, Guatemala, Colombia

Google, Antel (Uruguay), Angola Cables, Algar Telecom (Brazil)

2016

64 Tbps

10,556

Fortaleza, Santos

US

Junior

Google

2017

390

Rio de Janeiro, Santos

Seabras-1

Seaborn Networks

2017

72 Tbps

10,500

Fortaleza, Santos

US

South Atlantic Inter Link (SAIL), formerly CBCS

Camtel, China Unicom

2017

40 Tbps

5,900

Fortaleza

Cameroon

EllaLink

Telebras (Brazil), IslaLink (Spain)

2018

40 Tbps

N/A

Fortaleza, Santos

Portugal

Tannat

Google, Antel (Uruguay)

2018

90 Tbps

2.000

Santos

Uruguay

South Atlantic Angola Cables Cable System (SACS)

2017

40 Tbps

6,500

Fortaleza

Angola

BRUSA

2018

11,000

Fortaleza, Rio de Janeiro

US, Puerto Rico

Cable

Present owner(s)

Year

Capacity

Length (km)

South America 1 (SAm-1)

Telefonica (Spain)

2000

1.92 Tbps

South American Crossing (SAC) / Latin America Nautilus (LAN)

Level3, Telecom Italia Sparkle

2000

America Movil-1 (AMX-1)

America Movil

Monet

Telefonica (Spain)

Source: Elaboration by the authors based on various sources.

352


1990s: first generation of submarine optical cables It took only another six years after the commissioning of TAT-8 before the first of a new set of similar cables would connect Brazil to the rest of the world of optical communications. The first of these cables was Americas-1, that in 1994 connected Brazil to Florida with the arrival of the 500 Mbps-capacity cable in Fortaleza, where it connected to the terrestrial network of Embratel, still a state-owned company at that time. The Embratel network, already with national coverage, also included Florianopolis, where another submarine ca‑ ble, the Unisur, had also been laid in 1994, reaching Uruguay and Argentina, and connecting these countries to the US through Americas-1. In this same period, more than 2,500 km of submarine optical cables were laid along the the eastern coast of Brazil by Embratel in 1996, which would form the Brazilian Festoon system. In a festoon cable system, the underwa‑ ter cable stretches are relatively short, and require neither the installation of underwater amplification equipment, nor the use of electrical wiring within the optical cable, as a consequence. In this case the cables are passive and all transmission and amplification equipment is installed onshore. Figure 14.5 illustrates the Brazilian Festoon cable system of 1996, installed between the cities of Natal, in the state of Rio Grande do Norte (RN) and Santos, in the state of São Paulo (SP), including an onshore stretch between the south of the state of Bahia and north of the state of Espirito Santo.

FIGURE 14.5. Part of the national network of Embratel illustrating the festoon submarine cable system along the east coast of Brazil, built in 1996

Source: RNP (from Embratel).


353

Besides facilitating upkeep and operation, one of the additional advan‑ tages of a passive festoon cable system is the ease of upgrading the optical equipment in use, as it is not necessary to recover the cable from the bottom of the sea to replace this equipment. In the case of Embratel, the cable sys‑ tem that was laid in 1996 is still in operation, which is not the case of either Americas-1 or the Brazilian part of Unisur. According to information ob‑ tained from the manufacturer of the upgraded optical equipment installed onshore in 2014, the total capacity of the festoon cable has now reached 20 channels of 100 Gbps. Finally it is worth mentioning that, of these three initiatives, the two interna‑ tional ones were carried out by consortia of traditional telecommunications com‑ panies, many of them state-owned. They did not anticipate the future demands for international use of the commercial Internet that emerged in 1989 in the US.

The year 2000: full of novelties In the year 2000, no fewer than five new international submarine cable systems reaching Brazil were commissioned. We can distinguish between two categories of systems among these new rollouts. The first one includes Americas-2, a new cable with almost the same topology as Americas-1, and capacity of up to 7.5 Gbps, and Atlantis-2, following a new route connecting South America to Europe (Portugal), touching the coast of Africa, with 40 Gbps capacity. Compared with the previous cables, these figures represent great capacity increases. Nevertheless, these two systems did not and do not use the newly created technology of Dense Wavelength Division Multiplexing (DWDM), that allows on-demand provision of a great number (as many as a hundred or more) of communication channels in a single optical fiber, with‑ out altering the basic structure of the installed cable. The other three cable systems (SAC/LAN. SAm-1 and GlobeNet), all con‑ necting South America to the US, use DWDM technology, investing thus in the future expansion of its capacity along the useful life of the cable, and with total nominal transmission capacity in the range of terabits per second (Tbps), a hundred or a thousand times greater than those of Atlantis-2 and Americas-2. Another feature of these three systems is the adoption of redundant cable topology, so that the operator does not depend on using the cable system of another operator, potentially a competitor, to guarantee the continuity of its op‑ eration in case of connectivity failure, provoked, for instance, by damage to the submarine cable. Finally, very differently from joint cables (such as Americas-1,

354


Americas-2 and Atlantis-2), these three cables have few owners, frequently new business groups, whose origins do not stem from the period of state-owned telephone monopolies. Such groups were the first ones to recognize the poten‑ tial of investing in communications infrastructure which would prove essential for the explosive growth of the Internet at the end of the 1990s.7 The biggest of these large-capacity systems was originally owned by Global Crossing, which gave the name South America Crossing (SAC) to its cable system around South America. SAC performs a complete circuit of the South American continent, with onshore stretches in Panama, in the north, and between Argentina and Chile, in the south. This system has four pair of optical fibers, one of which was acquired by a company of the Telecom Italia group, that operates under the name Latin America Nautilus (LAN). Global Crossing was taken over in 2011 by the American operator Level3. All four fiber pairs of the SAC/LAN system operate with DWDM, with 10 Gbps channels, and together they have a 3.84 Tbps nominal capacity. The second system was installed by the Telefonica group from Spain and is called South America 1 (SAm-1). Its topology is similar to SAC, except that it crosses Guatemala instead of Panama. The cable uses two pairs of optical fiber, with 10 Gbps channels, and its nominal capacity is 1.92 Tbps. The third system, GlobeNet, consists of a ring entirely in the Atlantic Ocean and it serves mainly Brazil, Venezuela and Colombia, connecting them to the US. GlobeNet was taken over by Brasil Telecom in 2002, and passed to the control of Oi in 2009 when these two Brazil-based companies merged. More recently, GlobeNet was sold to the Brazilian group BTG Pactual. Its ca‑ ble uses four pairs of optical fiber, with 10 Gbps channels and total nominal capacity of 1.36 Tbps. The gamble on future growth in demand by the original owners of these three systems was justified, confirmed by fifteen years of their operation. As a result we may point out the concentration in Miami of Internet data flows between North and South America, elevating this city to the position of the second largest exchange hub of international traffic in the US, after New York, and the sixth one in the world. However, this also reflects the absence of al‑ ternatives to cater for international Internet traffic involving South America. 7 The historical evolution of the worldwide submarine cable network, affected by international politics, business organizations and technology, is described in the first chapter of Starosielski (2015).


355

2014: the arrival of transmission in multiples of 100 Gbps The practice of using 10 Gbps channels in DWDM systems began around the year 2000 and remained the de facto standard for more than ten years, even though 40 and 80 Gbps channels started to emerge in different optical systems. What determined the adoption of the new 100 Gbps standard was the standardization by the IEEE of 100G-Ethernet for data networks, recog‑ nizing thus the new primacy of data in communications networks. In our region, the first cable system to adopt the 100 Gbps standard, in 2014, was América Móvil 1 (AMX-1), from the Mexican business group that controls Embratel, Net and Claro in Brazil. More than a decade had elapsed without any large investments in new submarine capacity to Latin America. As AMX-1 received investment from América Móvil and involved such extensive coverage, encompassing a great portion of South America and additionally reaching the Caribbean, Central America and North America, the beginning of its operation gained great cov‑ erage in international media, as shown in Figure 14.6.

FIGURE 14.6. Infographic of the AMX1 optical underwater cable

Source: Adapted by RNP from Alcatel-Lucent.

Planned and foreseen cables for the coming years A string of projects of new submarine cable systems has already been announced for the next few years, with the first roll-out in 2016 and

356


across-the-board adoption of the new 100 Gbps channel standard. Of the projects listed in Table 14.1, one connects two points in Brazil, three connect Brazil to the US, one to Europe, one to Uruguay and two to Africa. As some of these projects involve similar routes (to the US and Africa) it is possible that not all of them will be carried out. Table 14.1 summarizes information on these projects, obtained from the Internet website TeleGeography,8 from which the cable topology of Figure 14.3 was also extracted. The following discussion is organized per route.

A. Brazil – US • Monet: the cable extends between Boca Raton, in Florida, and Santos, with partial access in Fortaleza, and will include six pairs of fibers, di‑ vided between the investors Google (2), Angola Cables (2), Antel (1) and Algar Telecom (1). The beginning of operations is planned for 2016. • Seabras-1: The cable extends between Brookhaven, New Jersey, and Santos and will include six pairs of fibers. Service is intended to kick off in 2017. • BRUSA: the cable extends between Virginia Beach, Virginia, and Rio de Janeiro, with access in Puerto Rico and Fortaleza. It will be built by the Spanish company Telefonica and its entry into operation is planned for 2018.

B. Santos to Uruguay and Rio de Janeiro (extensions of Monet) • Junior: The cable extends between Santos and Rio de Janeiro and will be built by Google. Operation starts in 2017. • Tannat: the cable extends between Santos and Maldonado, Uruguay, and will include six pairs of fibers. The investors include Google and Antel. Operations start in 2018.9

C. Brazil – Europe • EllaLink: the cable extends between Sines, Portugal, and Santos, with access in Fortaleza. In the future it will include points of access in 8

Available on: . Accessed: Apr. 26th, 2016.

9 Available on: . Accessed: Apr. 26th, 2016.


357

Cape Verde, Canary Islands (Spain) and Madeira (Portugal). Among its investors are Telebras and Islalink (Spain). It is intended to start operating in 2018.10

D. Fortaleza – Africa • South Atlantic Inter Link (SAIL), formerly called Cameroon-Brazil Cable System (CBCS): the cable extends between Fortaleza and Kribi, Cameroon. The investors are Camtel and China Unicom. It is intend‑ ed to start operating in 2017. • South Atlantic Cable System (SACS): the cable extends between Fortaleza and Luanda, Angola, and will have four pairs of fiber. The owner will be Angola Cables. It is intended to start operating in 2018. The new cables foreseen for the next two years, if they are built, will certain‑ ly have important impacts on Brazil. Firstly, three of the cables (from groups A and C above) will enormously increase the capacity available between Brazil and the Northern Hemisphere, in virtue of the adoption of 100 Gbps technol‑ ogy, ten times greater than the cables launched in the year 2000. This will certainly tend to drive down the cost of international communications. The other novelty will be the construction of transatlantic cables from Brazil to Europe and to Africa, opening new routes between Brazil and the rest of the world, without passing through US territory. This feature reduces communication latency (end-to-end delay), an important factor in global financial markets. In 2014, Edward Snowden’s revelations of digital espionage carried out by the US government caused Brazil’s government to seek new international telecommunications routes, avoiding the US, to lower the risk that information would be intercepted in transit. Finally, the new cables will provide the global telecommunications network a higher degree of connectivity, with many new alternative routes, compared to the current topology. For these reasons we see great benefits for Brazil that will come out of the new proposals.

10 In 2015, the name of this cable was changed to EllaLink, to be built by the joint venture that bears the same name. Its RFS (Ready for Service) deadline was put off until 2018.

358


SUBFLUVIAL CABLES IN THE AMAZON REGION – THE AMAZONIA CONNECTED PROGRAM The existing long-distance optical infrastructure in the Amazon region In December 2015, the Brazilian Amazon region could count on the following long-distance optical routes (Figure 14.7): • Porto Velho – Manaus: Aerial optical cable from Embratel, suspend‑ ed from poles built along a federal highway, BR319, that crosses the dense forest zone between these two state capitals, complemented by a subfluvial cable crossing of the Amazon River at Manaus. • Tucuruí – Macapá – Manaus: Aerial OPGW (OPtical Ground Wire) system from TIM and Vivo, using the electric power transmission lines that connect these two capitals to the Tucuruí hydroelectric pow‑ er plant in the state of Pará (PA). • Belém – Tucuruí – Marabá – Altamira – Santarém – Itaituba: Aerial OPGW system from Eletronorte, using the electric power transmission lines that connect these cities to the Tucuruí hydroelectric power plant. • Manaus – Coari: Ducted optical cable from Petrobras, built along the route of its gas pipeline between Manaus and Urucu, in the state of Amazonas (AM), with a subfluvial crossing of the Solimões River near Coari. • Manaus – Boa Vista: Ducted optical cable from Oi, buried along a federal highway, BR 174, crossing indigenous areas between these two state capitals. • Macapá – Oiapoque: Ducted optical cable from Oi, buried along a fed‑ eral highway, BR 156, through an sparsely populated area, reaching the border with French Guiana.


359

FIGURE 14.7. Map of the long-distance optical routes in the Brazilian Amazon in 2013

Source: RNP.

Though important for the Amazon region, these optical routes intercon‑ nect almost exclusively the endpoints, not providing service to a large part of the population, which does not live along these routes, but more particularly along the courses of the major rivers of the region.

The use of subfluvial cables in the Amazon region For many years terrestrial optical fiber routes have been built along roads, railroads and electric power transmission lines. These routes have been used to offer support to the national and regional telecommunications network backbones, especially Internet transport backbones. Moreover, as previously described, submarine optical cables have been used extensively for the in‑ tercontinental interconnexion of national and regional telecommunications networks, giving them worldwide reach. Nevertheless, in many countries, there are still regions where there are no such terrestrial routes, notably in regions of difficult land access and low population density. Some such regions in the world are crossed by rivers that provide a common element to meet the countless needs of society. One such region is Amazonia, where there are a few roads serving the main cities and where the population inhabits riverside zones along its great rivers that cross

360


the region and provide both food and the main means of transportation – river transportation. In such regions, one should consider the alternative of installing subfluvial cables along the courses of the rivers. Besides providing communications and Internet services to the riverside populations, a route along the river causes much less environmental damage than the construc‑ tion of a land route that cuts through the forest.

The Amazônia Conectada [Connected Amazon] Program The Connected Amazon program [Programa Amazônia Conectada] was launched in 2015 by the Brazilian Army, which has led its gradual rollout, in partnership with the National Education and Research Network (RNP), the state IT companies Prodam (the Amazonas state ICT company) and Prodepa (the Pará state ICT company), the governments of the states of Amazonas and Pará, the federal government-owned telephone and electrical energy compa‑ nies (Telebras and Eletrobras), among others [PROGRAMA AMAZÔNIA CONECTADA, 2015]. The program’s objective is the construction of a subfluvial optical cable mesh along the courses of the main rivers of the Amazon region, where there are few roads or railroads or even electric power transmission lines. The program seeks to align itself with the federal government’s National Broadband Program (Programa Nacional de Banda Larga – PNBL, discussed in Chapter 3 of this book). More specifically, the new optical cables in the rivers will provide terrestrial access to telecommunications and Internet services to the riverside populations, thus contributing to the digital inclusion objectives of the PNBL. Further alignment is expected with the state governments that are building out their own statewide communications networks. The seed of this program was the proposal put forward by RNP in 2013 of the rollout of the Belém-Macapá-Manaus subfluvial route, closely following the route used by the telegraph cable deployed between 1895 and 1896, pre‑ viously mentioned in this chapter (SIEMENS, 1896). This proposal was pre‑ sented initially to the then Ministry of Communications (MC) as a challenge to improve the telecommunications infrastructure of the region, extending the availability of broadband services and of communication in general, as well as bringing benefits to a large part of the population in the region that lives mostly along its rivers. The MC itself counterproposed the idea of deploying not only along this route, but instead building a much larger mesh of optical cable routes to be


361

installed along the major rivers of the Amazon region – the Amazon, Madeira, Negro and Solimões Rivers (GRIZENDI; STANTON, 2013). However it was the Brazilian Army that responded to this challenge and proposed the cre‑ ation of the Amazonia Connected Program. The main benefits of this program are: • To build a telecommunications infrastructure in the Amazon re‑ gion that will boost regional development and add to the strength of National Defense and Public Safety; • To establish a solid foundation for the development of research and education networks in the North of Brazil; • To contribute to Brazil’s technological and industrial development, with global scale and competitiveness The program has an estimated extension of 7,800 km and encompasses five routes called Information Highways [Infovias] along the following rivers: Negro, Solimões, Madeira, Purus and Juruá, as shown below (Figure 14.8).

FIGURE 14.8. Planned subfluvial cable routes in the Amazonia Connected Program

Source: RNP

362


FIGURE 14.9. Subfluvial cable route between Coari and Tefé (in detail)

Source: RNP

The first Information Highway to be built is along the Solimões River, between Manaus and Tabatinga, on the international border with Colombia and Peru. Initially, it was decided to begin this route in Coari, rather than Manaus, taking advantage of the existing Petrobras fiber be‑ tween Manaus and Coari. Thus, the first subfluvial cable stretch upstream from Coari – effectively, the pilot project – interconnects Coari and Tefé, 240 km apart, and this was concluded in the first semester of 2016. Figure 14.9 illustrates this route. Even earlier, before carrying out this pilot project, a smaller “technolo‑ gy demonstration” project was successfully carried out in the Negro River at Manaus, where a stretch of about 7 km of subfluvial cable was laid, intercon‑ necting two Brazilian Army units and completing an optical ring with RNP’s Metropolitan Network of Manaus – MetroMAO (see Figure 14.10). This earli‑ er project was publicly demonstrated on July 16th, 2015, when the Amazonia Connected program was formally launched through the Interministerial Ordinance n. 586, signed by the Ministers of Defense, Communications, and Science, Technology and Innovation.


363

Laying and operating 7 km of cable in the Negro River was relatively sim‑ ple. The pilot project along the Solimões River may present bigger difficulties. The report of the deployment of the telegraph cable along the Amazon river (SIEMENS, 1896) mentions problems of cable upkeep during the rainy sea‑ son, when the combination of strong currents and debris in suspension led to the scouring of the riverbed, leaving the cable uncovered and damaged. The great challenge will be to find a way of adequately burying the cable in the bed of the river while it is being laid.

FIGURE 14.10. Technology Demonstration project with the laying of about 7 km of subfluvial cable in the Negro river in Manaus

Source: RNP

CONCLUSIONS The deployment of this initial stretch along the Solimões River already represents something quite unusual in modern telecommunications. If one considers that it was carried out in the Amazon region, one must recognize that it is a really challenging task, that can bring technological advances and innovation through its extension to several of the big rivers of the region. Compared with submarine cables, subfluvial cables are simpler, since they are used in a festoon topology, with transmission equipment located on dry land. On the other hand, the strong currents and the frequent shallows turn the rivers into more troubled environments, compared with the tranquility

364


found in the bottom of the oceans, potentially making both the deployment and corrective maintenance of subfluvial cables more complex. Finally, most of the population of this region lives along the courses of these rivers, which already provide them with transport and food, and it seems to us quite un‑ precendented for these same rivers also to provide them with access to 21st Century communications technologies. When the Connected Amazon Program is complete, it will bring high-quality high-bandwidth communication to a region that is currently very poorly served both in the capacity and quality required for basic tele‑ communications services. The Coari – Tefé project will be valid as a proof of concept so that the solution employed may be extended in the whole of the Amazon region, by overcoming technological challenges and creating a knowledge base for projects in other stretches of the Solimões River, as well as for projects in other rivers in the Amazon region and even in rivers of other regions in the country. The mastery of this technology also should open up opportunities for Brazilian submarine and subfluvial optical system installation companies in other countries and continents, with special attention to neighbouring countries of the Amazon region, such as Peru and Colombia, that share the same geographic, climatic and vegetation conditions as Brazil. In the case of Peru, a project is being elaborated, with the participation of a Brazilian company, to extend the land-based national telecommunications network to reach the city of Iquitos, in the heart of Peruvian Amazonia. A survey has already been undertaken for the deployment of a 450-km subfluvial cable between Nauta and Yurimaguas along the Marañón and Huallaga Rivers, tributaries of the (Peruvian) Amazon River, which is the upstream continuation inside Peru of the Solimões River. A further possibility is to deploy a 420-km subfluvial connection between Iquitos and Tabatinga, in the Brazilian state of Amazonas (AM) on the border of Colombia and Peru. The interconnection of these two cables along the Solimões, (Peruvian) Amazon, Marañón and Huallaga rivers would make possible a direct con‑ nection between Peru and Brazil, consequently bringing benefits to inter‑ national telecommunications (Figure 14.11).


365

FIGURE 14.11. Possible subfluvial cable interconnection routes in the Peruvian Amazon region

Source: RNP Note: The smaller map also shows the IIRSA Norte [North IIRSA] highway, newly built between Yurimaguas and the port of Paita on the Pacific Ocean.

Other possible candidates for adopting the solution used in the Amazonia Connected Program would be countries with large rivers and dense rainfor‑ ests that are located mainly in Africa and Asia, especially in the basins of the Congo (or Zaire) River in Africa and of the Mekong River, in Southeast Asia. Brazilian companies’ expertise with the approach used in the Amazon could be applied in such cases. It should be noted that both new cabling projects with large or majority Brazilian participation, EllaLink and Amazonia Connected, depend greatly on investments by the Brazilian government. In both cases, the reasons for building these cable systems involve public policies of meeting the needs of broadband communication of the Brazilian population, with differen‑ tiated international service (access to the world, without passing through the US, and with lower latency). At present there seems to be no interest on the part of the Brazilian private sector in investing in these initiatives. Part of this problem may be explained by the fact that most of the large

366


telecommunications companies operating in the country are under the con‑ trol of international business groups, with financial interests in other cable systems competing with EllaLink, which, however, pass through the US or, in the future, through Africa.

REFERENCES PROGRAMA AMAZÔNIA CONECTADA. Available on: < http://www. amazoniaconectada.eb.mil.br/eng/ >. Accessed: October 30th, 2016. BRIGHT, C. Submarine Telegraphs: Their history, construction and working. London: Crosby Lockwood and Soon, 1898. Available: . Accessed: Oct. 4th, 2015. COSTA DA SILVA, M. A telegrafia elétrica no Brasil Império: ciência e política na expansão da comunicação [Electric telegraphy in Imperial Brazil: science and politics in the expansion of communication]. Revista Brasileira de História da Ciência [Brazilian Journal of the History of Science], Rio de Janeiro, v. 4, n. 1, p. 49-65, Jan./Jun. 2011. Available on: . Accessed: Mar 7th, 2016. EASTERN TELEGRAPH. Available on: . Accessed: October 6, 2016. FRANK LESLIE’S ILLUSTRATED NEWSPAPER, Aug. 21st, 1858. Available on: . Accessed: September 19th, 2016. GRIZENDI, E.; STANTON, M. A. Use of subfluvial optical cable in a region without landbased infrastructure: a project to deploy optical cable in the Amazon region. Proceedings and Reports of the 6th UbuntuNet Alliance Annual Conference, Kigale, Ruanda, p. 5368, 2013. Available on: . Accessed: November 2nd, 2015. SIEMENS, A. Cable laying on the Amazon River. In: The Electrical Engineer, 1896. Available on: . Accessed: October 1st, 2015. STANDAGE, T. The Victorian Internet. Londres: W&N, 1998. STAROSIELSKI, N. The Undersea Network. Durham: Duke University Press, 2015. SUBMARINE CABLE MAP. Available on: . Accessed: March 25th, 2016.

| 15 |

FIBER IN THE FAVELA: OUR EXPERIENCE IN ROCINHA Asafe Coimbra Samuel Silva [ Net Rocinha ]

I

INTRODUCTION

n this chapter we will introduce a bit of the story of Rocinha, a commu‑ nity in Rio de Janeiro located in the southern part of the city, close to up‑ per-middle class neighborhoods. There, in the late 1990s, we decided to venture out as entrepreneurs, first in computer services and later as Internet providers, though we had not yet met. In 2011, with the warnings about the entry of the Pacifying Police Unit (Unidade de Polícia Pacificadora – UPP), we met several times with local providers. It was during one of these meetings that we met and started our business partnership. Let us tell you a little of our story and that of the Net Rocinha, our compa‑ ny, how it emerged and grew, overcoming the challenges it faced. There is no victory without a struggle!

ROCINHA In the 20’s Rocinha was a farm that began to be occupied and subdivid‑ ed into smaller plots, which soon became the main suppliers of food for the “Gávea street market”. The region was becoming increasingly populated, with disorderly growth, which resulted in the creation of a shantytown (known in Rio de Janeiro as a favela), today a community, home to about 200,000 inhabitants in 2016, though only 69,161 residents were enumerated by the Brazilian Institute for Geography and Statistics in the 2010 census.1 1 Available at: . Accessed: January 12. 2016.

368

asafe coimbra • samuel silva

With such a large number of people living in such a small area, the de‑ mand for essential services became increasingly evident. So, these services started to be deployed: local shops, public transport, hospitals, power stations, telephone booths, public schools, nurseries, sports education centers. In addition to these services, neighborhood associations started to emerge. Already in the 1990s, the community started having access to technology, bringing a telephone company to provide Internet service that was still very expensive and inefficient.

SAMUEL’S STORY In 1998, I finished a specialization course in computer sciences, while working for a motorcycle taxi service. I started my activity in technology serv‑ ing homes, repairing computers. In 2004, I decided to open my first comput‑ er company. Being an essential tool for the computer services, I soon realized the lack of access to the Internet was a very big obstacle. At that time, only one operator –Telemar (currently Oi) – offered Internet access service in the Rocinha. So I went on, despite not yet being regularized. I managed to obtain a considerable income and was then saving up money. Always first seeking to provide quality services; and thus I managed to gain the trust of customers and began retaining the majority of them, because as soon as they purchased computers, they realized that without Internet, com‑ puters weren’t very useful, that is, they faced the same needs and the same obstacles I encountered. As a customer of Telemar, I could see that that operator was not providing a good service. I can say that its outdated and abandoned network in Rocinha made providing quality service impossible. Today, it is still pretty much like that. These conditions have opened the door to the growth of Net Rocinha. I undertook a market study and I realized that there was need for access to the Internet that would extend beyond my loyal customers. I then started to look for technical and regulatory information on how to become a provider, and came to the conclusion that it was quite difficult and bureaucratic to legalize an Internet access provider. I was, however, already certain that it was a great niche to be exploited and decided to start providing it, even clandestinely. The first customers were those for whom I was already doing maintenance work on their computers. These customers referred my

fiber in the favela

369

services to their friends saying they were getting a good service, and in three months I had already had 100 new customers. The difficulties were many. I started this work with no funds for in‑ vestment, using part of my working capital to buy equipment and basic cables, that is, I used household products and still managed to serve my clients with a better quality than the incumbent carrier. Afterwards, I start‑ ed purchasing more professional equipment, suitable for the Internet and telecommunications. At that time, I catered to customers in a “hidden” office located in one of the several community alleys and even used my personal mobile number for customer support. In 2008, I rented office space on Rocinha’s main street and people came to know me better. That was when I started to meet my competitors. In 2008, the Morro Dona Marta was the first favela to be pacified in Rio de Janeiro.2 Soon there were rumors that Rocinha would also be “pacified”, but almost no one believed it. With this imminent occupation of Rocinha by the UPP, I intensified my search for information on how to legalize my business. Then an unprecedented event occurred at Rocinha: all Internet pro‑ viders decided to meet to become regularized. Nobody knew anything about regulation and about the General Telecommunications Law (Lei Gertal de Telecomunicações – LGT), so we hired a telecommunications engineer to help us. His first suggestion was to try to set up a cooperative scheme so that every‑ one would join and obtain a Multimedia Communication Service (Serviço de Comunicação Multimídia -SCM) license. A politician and a community leader became interested in our cause and tried to help, but I did not believe that this idea would work. In one of those meetings I noticed, “there in the corner,” a well-spoken guy who was knowl‑ edgeable about technical issues, I approached and invited him to join me in a partnership. He did not believe that a cooperative would work either, so he soon became interested in collaborating and we began to meet separately from the rest of the group. Soon after that we left the idea of the cooperative behind and started our partnership.

2

Pacification here means the process of expelling or at least disarming drug dealers and other criminals and installing a UPP manned by specially trained police officers.

370


FIGURE 15.1. Asafe Coimbra and Samuel Silva in Rocinha

Source: Author’s personal files

ASAFE’S STORY My first job was in a company located in the Gavea district in 1997. There I made deliveries and repaired various electrical appliances, gaining good experience in this area. In 1999, a friend recommended me to work in an electronics repair company in Ipanema, in a shopping mall, and there I also made deliveries and small repairs, always supervised by a technical manager. Among others appliances, I had a sense of how televisions, stereos and cord‑ less phones worked. This experience was excellent. In this same shopping mall, I ran across a computer company and went to work for it, seeking to learn everything I could about computers. I held this job for only six months, but it was enough to learn how a computer worked. I learned about hardware, how to assemble computers and install operating systems and programs; and at the time Windows 98 was the most used OS. Good times! Close to this computer store, across the street, there was a musical instru‑ ment store, the famous Space Music, today known as “Crazy for Music” (Louco Por Música). I started working there. After some time, I became responsible for the store and also learned to repair musical instruments such as guitars

fiber in the favela

371

and electric guitars. Then, with the change of store address in 2005, I started working in Copacabana for Louco por Música. It was my last job before open‑ ing my own business Still while I was working at Louco por Música in mid-2006, and since I loved computers, I began to see a need in the community where I was born and still live, Rocinha, since there was no Internet, and very few families had a computer at home. Seeing this, I asked a friend if I could borrow his Velox line (Oi’s broadband service). So I began to distribute Internet informally (we call it “gatonet”) in the upper part of Rocinha. Soon I was getting a growing number of Internet users as customers. Even today it is very difficult to get a landline in Rocinha, and back then it was even harder. However, I insisted with various friends and one of them sold me a line. It was a start for the business to grow and spread easily, in‑ creasing the number of customers and Internet capacity. At that time, I had around 50 customers. Since I could no longer provide support to my clients and also work at Louco por Música, I decided to quit and work on my own. I left Louco por Música in January 2007 and in February rented a store located closer to the Rocinha’s main street. There I set up an Internet cafe (known as a lan house in Brazil), starting with three computers bought with my savings from Louco por Música. Later I bought more computers with the revenue from my infor‑ mal Internet service provider (ISP) to residents. I’m very proud of having been the one to introduce computers in the upper part of Rocinha: offering kids access to the Internet was a great thing for them, creating Orkut (a social network) accounts, promoting gaming tournaments, such as CounterStrike, among others. Those were very good times. As time passed, the lan house came to be open around the clock. That was too tiring for me. That is why I decided to close the lan house and open a com‑ puter store, selling products and services, including assembly of computers and accessories such as mice, speakers and keyboards. The store also offered maintenance service for customers’ computers and laptops. At that time, as a provider, I started looking for partnerships with com‑ panies already having an Anatel SCM license, seeking to legalize my ISP business. It was in 2010, with the announcements of the imminent arrival of the UPP in Rocinha (the UPP was only opened in September 2012), all local

372


suppliers such as myself met to create a cooperative to solve all our prob‑ lems and legalize our companies. We held a variety of meetings and in one of those at the “House of Peace” Casa da Paz, I met Samuel. We started talking, getting to know each other and discussed various matters, which in my case were technical issues. I saw that Samuel had organizational ability and team management skills.

MERGER OF OUR TWO COMPANIES TO CREATE NET ROCINHA We realized that if we merged our businesses, we would be much stron‑ ger. We decided to abandon the idea of the cooperative and set out to create Net Rocinha. We met some partners in other locations in Rio de Janeiro who were hav‑ ing the same difficulties and also who did not believe a cooperative would work. We met and started discussing the idea of setting up a major company with five providers. But it did not work out. However, we have maintained our friendships. At that time, we already had contact with an engineer who offered consulting services to the cooperative group. Though we set out in another direction, we ended up hiring him to help us get our SCM license. We faced many problems, since we knew practically nothing about the SCM procedures. The engineer took advantage of this situation to make money off us without solving our problems. It took 15 months from the time we opened the company to the publication of our Multimedia Communication Services (SCM) license in the Official Gazette, with half of that time being lost in the Regional Council of Engineering and Agronomy (Conselho Regional de Engenharia e Agronomia – CREA-RJ). While the SCM licensing process was underway, we invested heavily in technology: fiber optics deployment, acquisition of servers and high-per‑ formance equipment, as well as professional training courses for all staff (partners and employees). We also invested in robust management and network monitoring systems, an Autonomous System Number (ASN), IPv4 and IPv6 address blocks and routing protocols necessary to have bet‑ ter performance, i.e., Border Gateway Protocol (BGP) and Open Shortest Path First (OSPF).

fiber in the favela

373

FIGURE 15.2. Our team in the Net Rocinha office


In the midst of our difficult struggles to improve quality, performance, service and legality, the Civil Police’s visited one of our stores, requesting documentation for legal provision of Internet service and the entire set of documents for any regular company. We had retained a company of that same consultant that was obtaining our SCM, a Value Added Service (Serviço de Valor Adicionado – SVA) partnership that allowed us to offer Internet service on behalf of another company while our license was being obtained. For a simple Portuguese language error in the contract we ended up in the police station to provide further clarifications. We were treated as “delinquents” and forced to go to the police station in a police car. In the end, everything worked out and the police chief realized it was only a misunderstanding, but it was a very embarrassing situation. It is not easy to provide Internet service inside a community. Peculiar diffi‑ culties force us to develop special techniques, for example, make network con‑ nections and carry out maintenance services in alleys and narrow streets. The problems we faced were (and still are) many: saturated poles with dozens and perhaps hundreds of anchored cables, illegal electric power connections that pro‑ voke oscillations in the network and cause constant fires, even more frequent‑ ly during the summer (Figure 15.3). Eighty percent of our cabling is located on

374


narrow streets and alleys, which prevents the use of vehicles to reach the custom‑ er. Rocinha has heavy traffic and this complicates the use of vehicles even on the main streets. The risk of death is imminent because of constant shootings (in fact, before the UPP that risk was not as big) and theft of equipment and external cables is a daily risk. Even our fiber cables suffer from gunfire.

FIGURE 15.3. Cable-saturated poles at Rocinha


Among the many difficulties, we still have a hard time hiring employees. In addition to the lack of skilled labor in telecommunications in our commu‑ nity, we have the problem of prejudice from professionals who live in other neighborhoods and are not willing to work in our company for fear of the violence shown in the media. We have only a few successes in hiring from the “outside”. We suffer prejudice from some banks, that do not see the potential that there is in our company, and they end up refusing to make loans to us. We know that, in fact, the company’s and the partner’s addresses, being in a favela in Rio de Janeiro, have a negative influence on the banks’ decisions.

fiber in the favela

375

Having numbering resources (ASN), IP blocks, fiber-optics, good network traffic engineering, redundant routes of dedicated fiber link, connection to the PTT-RJ and PTT-SP (Internet Exchanges), network monitoring, personalized ser‑ vices (Facebook, WhatsApp, 0800, shop, online chat, relationship center), Anatel license, competent staff, payment of taxes, high speed plans for a fair price (20 Mbps + WIFI for R$ 89.90, that is about US$ 36) makes Net Rocinha the most qualified Internet company among those that offer these services in Rocinha. Amid so many obstacles overcome, we still have our much-valued so‑ cial responsibility program, as you can see in our Fan page on Facebook. com. We support some social projects and among them, a special one that transforms the lives of Rocinha’s kids through sports. Managed by “Uncle Ley,” we have a project that currently involves over 100 children, teaching them bodyboarding. In the same project we have an athlete to whom we offer great support, David Barbosa, a 17-year-old boy who is super-talented, participating in world championships in the professional category and who has even competed in Chile.

FIGURE 15.4. Uncle Ley and his bodyboarding students on São Conrado beach


376


We just want to do good. Helping is part of our mission and today, some public nurseries at Rocinha, which had no access to technology, receive our Internet service. We are sure that teachers need access to information and we are helping them with that. We know that this support will be reflected in the future of children studying in these nurseries. We have the project to bring the Internet to all community nurseries and this should happen soon. We are the first provider legalized by Anatel in Rocinha. We are the only company in Rocinha receiving an ASN from Registro.br, which allows us to access a block of IP numbers to deliver to our clientes.3 We pioneered in the use of optical fiber in the community, doing traffic engineering, with network sectors and improving the quality of Internet access. The quality of our service is our advantage, because we sell and deliver the contracted speed. Today our customers can surf the net at 20 Mbps, quickly accessing all sites, including games, with the possibility of having a fixed IP number and network monitoring, which may be common for a major carrier, but here in Rocinha, no other provider offers this ser‑ Table 15.1. vice. Often operators sell, but do not deliver what they Number of sell, deceiving consumers. customers per year We are now connected to Rio de Janeiro’s PTT, which was another major step. We continue working Clients Year served to improve our ability both in manpower and technol‑ ogy. We feel proud to be part of this family called Net 2008 150 Rocinha. We had 1,600 customers at the end of 2015 2009 200 and our goal is to have 4,000 by 2017 (Table 15.1). 2010 300 Technology advances at a very fast pace and we 2011 500 are always ready to grow. But the path to growth 2012 700 would be easier with funding opportunities from 2013 1,000 partners who believe in our potential. We analyzed fi‑ ber to the home (FTTH) three years ago and we have 2014 1,400 no doubt that such a network topology will dominate 2015 1,600 the telecommunications market. To follow the mar‑ 2,800 2016a ket, we are doing everything possible to bring fiber to 4,000 2017a the subscribers’ homes. With the use of optical fiber Note: in our backbone we already managed to see a great a Goals reduction in problems and with the deployment of 3

For further information on Registro.br, the reader may visit http://registro.br. See also Chapter 10 of this book.

fiber in the favela

377

FTTH, we expect to achieve a reduction of up to 80% in problems and a growth of at least 50% per year. For example, before the deployment of fiber, the network originated from a single point, with many flaws and constant repairs. But with the deployment of fiber, we divided the service into sectors, offering more quality. The market is promising in the region.

AT ABRINT – LEARNING AND REPRESENTING Amid the growth and legalization of Net Rocinha, we got to know the Brazilian Association of Internet and Telecommunications Providers (Associação Brasileira de Provedores de Internet e Telecomunicações – Abrint – see Chapter 11), which plays a key role as our partner. Our first contact with Abrint was in 2010, during the 2nd Meeting of Providers in São Paulo. In this meeting we got a better understanding on how our sector functions. We met many people with extensive experience in information technology and telecommunications, and felt like we were in a school. We joined Abrint and started talking about our company and sharing our experiences, and along the way we made friends all over Brazil. We managed to fulfill many obligations and complied with formal re‑ quirements of a telecommunications company with the help of colleagues we met through Abrint. One example was the issuing of model 21 (communica‑ tion) invoices that not even our accountant knew how to issue. We shared our story with all the members and followers of the Association. Shortly after we joined the Association we received the prestigious visit of Breno Vale and several other Abrint directors, so they could get to know our reality up close. A reporter and a photographer of the O Globo newspaper in‑ terviewed us, taking pictures and recording videos on the difficulties we had at that time, and what we did to solve them. Every day that went by we felt closer to Abrint and so Samuel decided to apply for the position of member of the Advisory Council. He had nev‑ er thought of it and, to his surprise, he had 79 of the 120 valid votes in the Annual General Meeting held on June 1, 2015, during the 7th National Internet Providers Meeting. This result makes us believe that our history is truly of great value and can help many other providers with the same difficulties.

378


PUBLIC POLICIES TO PROMOTE THE DEVELOPMENT OF BUSINESSES SUCH AS OURS The biggest villain for a new entrepreneur in the telecommunications in‑ dustry is the high cost of purchasing equipment and of network deployment. We see a constant struggle of several friends in the industry: financial institu‑ tions do not understand the nature of small and medium-sized Internet and telecommunications service providers. For example, banks want a guarantee when they offer loans, but the optical fiber cables that we purchase are not suitable for this purpose: in case of default, the bank has no means of taking these cables. In 2015, however, Abrint began talks with the Ministry of Communications and Anatel, seeking to establish a credit guarantee fund for our category. If this really happens, it will be a major step facilitating the growth of our busi‑ ness and the development of our country. Another priority for public policy is to streamline and simplify the re‑ quirements for the legalization of small and medium-sized telecommunica‑ tions companies. For example, Net Rocinha took 15 months to receive its SCM license, and many other small businesses face similar difficulties to legalize their busi‑ ness. Without formalization there is no way of investing and growing. In a community such as Rocinha, we also find it important that all pro‑ viders collaborate in the cleaning and organization of the cabling on poles. For example, when a cable is broken, the operator often does not reuse the cable, which results in many dead cables on a given pole. We suggested a way to identify the broken cables, which are often on sidewalks, and request an authorization from the operator so that we may clean these poles and reuse the cables.

PREPARING FOR THE FUTURE BASED ON ISO 9001 In the second half of 2015, we hired a consultant to help us structure our company, training managers and employees in all areas, preparing it to meet new challenges. This consultant has brought a new outlook and more professionalism to all of us, with the implementation of Business Process Management (Business Management with Process Mapping) in addition to all other controls necessary for the company, such as financial management, cost and price of sales management; human resources management, 5S

fiber in the favela

379

program, balance sheet analysis and financial statements, plus several other controls necessary for us to structure, empower ourselves and have continu‑ ous improvement as our standard, to better serve our customers every day. We still have a lot to learn, but we can say with conviction, that it was one of the best investments we have currently made. We realized that we need to invest not only in technology but also in professional learning, personal devel‑ opment as entrepreneurs and training of all staff so that we can grow, giving us sustainability, knowledge and experience to achieve our goals. With the consulting effort we also we had the opportunity to add more special software to control our fiber optic network. That has enabled us to know where each terminal is located, and all the places where we have cus‑ tomers, always operationalizing and streamlining our technology skills and increasing our credibility with our customers, in addition to making signifi‑ cant improvements to the services we provide.

CONCLUSIONS We were born and grew up in Rocinha, a community (previously a favela) with a prime location, which suffers from the violence of drug trafficking and the lack of telecommunications infrastructure, in addition to other ad‑ versities. We overcame many obstacles creating our company without hav‑ ing any college education, but with determination and will to overcome any obstacles we faced and that still persist. We were able to learn the basics of information technology and telecommunications to offer services better than those of the major carriers and capture more customers, and also con‑ tribute towards the development of our community, now even more with this business management consulting. We hope that our example will inspire other small and medium-sized Internet service providers and telecommunications to follow this path, pro‑ moting the development of our country. Yes, we can do it!

| 16 |

Conclusions Flavio Feferman [ University of California, Berkeley ]

O

ver the four parts of this book, the authors discuss the past, pres‑ ent and possible future of broadband Internet in Brazil. They make it clear that Brazil must follow the global progress of broadband Internet, while ensuring access to all Brazilians. As many authors point out, broadband Internet is increasingly important for economic, social and human development. The network also contributes to more inclusive citizenship, as the Internet facilitates access to a wide range of public services and serves as a means of expression in a democratic society. Several authors emphasize the role of broadband Internet as an essential service, much like access to education, health, electricity safe water, and sew‑ erage. Other authors underscore the role of broadband Internet in promoting productivity and competitiveness. As João Moura summarizes in Chapter 12: The digital economy serves as a fundamental enabler for other sec‑ tors that need to move forward in the country. Education, public services, industry, professional services and many other sectors to‑ day require access to 21st century infrastructure and next generation networks to thrive. New business models created by the Internet cannot move forward in Brazil without this resource.

New content and applications – including video, cloud computing and the Internet of Things (IoT) – require increasingly advanced telecommunica‑ tions infrastructure. At the same time, the availability of new infrastructure and telecommunications technologies, such as 100 Gibabit networks and 4G mobile access (or 5G in the future), unlock opportunities for new services and business models. In Chapter 2, Eduardo Tude provides definitions and important concepts related to broadband, and emphasizes the importance of both fixed and mobile technologies for expanding access to broadband

382

flavio feferman

Internet. Therefore, infrastructure expansion propels the evolution of con‑ tent, and vice versa, in an iterative and continuous process. This dynamic implies that the concept of “broadband” evolves over time. In 2015, the Federal Communications Commission (FCC), the regulatory agency of the telecommunications industry in the United States, raised its broadband Internet definition by 500% – from about 5 megabits per second (Mbps) to 25 Mbps. Cisco, a network infrastructure company, estimates that the average Internet speed in United States and Europe should almost double in the coming years, from 21.8 Mbps in 2014 to 40 Mbps in 2019. However, Cisco also predicts that the average speed in Latin America, Asia and Africa should remain far below this level (with some exceptions). Therefore, in the absence of public investments and policies, the “digital divide” between de‑ veloped countries and developing nations should increase.1 This digital divide is even more important within a country of continental dimensions and with great regional and social inequalities like Brazil. The chapters in this volume address a number of issues related to the expansion of infrastructure and access to broadband Internet in Brazil. The different perspectives presented make it clear that the evolution of broadband in the country requires the participation of various stakeholders: government and regulatory agencies, the private sector (small, medium and large com‑ panies), academia and civil society. Each has an important role to play, but each stakeholder group alone cannot fully address the challenges related to infrastructure expansion and access. Indeed, some of the most interesting cases presented in the book involve cross-sectoral collaborations: such as the example of the Ceará Digital Beltway (Cinturão Digital do Ceará, CDC), dis‑ cussed in Chapter 13 by Carvalho, Feferman, Knight and Woroch; the projects developed by the National Education and Research Network (Rede Nacional de Ensino e Pesquisa – RNP), presented in Chapter 9 by Eduardo Grizendi and Michael Stanton; the Connected Amazon program discussed in Chapter 14 by Grizendi, Stanton and Decílio Sales; or even the dialogue that began between the federal government and Abrint, an association that represents small and medium sized telecommunications companies, as explained by Basilio Perez and Breno Vale in Chapter 11. As Perez and Vale point out, the Multimedia Communications Service (Serviço de Comunicação Multimídia 1

HARDY, Quentin. The World’s Coming Broadband Divide. The New York Times, May 31, 2015. Available at: . Accessed: February 26th, 2016.

conclusions

383

– SCM) regulation established by the government enabled the rapid growth of small regional Internet service providers (ISPs) in the country. Throughout this book, the authors highlight six fundamental factors for the expansion, quality improvement, and “universalization” of broadband Internet in the country, summarized below:

1. SECTORAL REGULATION FOR COMPETITION AND INNOVATION Several authors – including Balbino, Tude, Moura, Perez and Vale, Foditsch and Belli, Alimonti and Lefèvre – highlight the importance of reg‑ ulation to promote competition, investment and innovation. This includes sectoral policies to encourage new investments and to promote the entry of new companies in the market, including: infrastructure sharing (un‑ bundling, among other forms of sharing), regulatory models for spectrum sharing (for wireless broadband), regulation related to companies with sig‑ nificant market power, improved access to financing (including credit guar‑ antee funds), as well as procedures to facilitate obtaining operating licenses for small and medium companies. In Chapter 4, Abraão Balbino introduces the concepts of “broadband input chain” and the “investment ladder”, emphasizing the importance of reducing entry barriers for new companies, by means of infrastructure shar‑ ing along the input chain. This sharing allows new entrants to carry out investments at other levels of the input chain, while bringing innovations to the market. Balbino conducts a detailed analysis of the General Program for Competition Targets, (Programa Geral de Metas de Competição, PGMC), which aims to increase competition and consequently influence the price and quality of broadband services. Veridiana Alimonti (Chapter 3) and other authors such as Moura, Barbosa, and Perez and Vale draw attention to the disparities in access and the lack of competition in broadband services within Brazil, especially outside large cit‑ ies. Alexandre Barbosa and his co-authors (Chapter 6) utilize CGI.br’s house‑ hold survey data to illustrate the considerable differences in access among the geographical regions of Brazil, between urban and rural areas and across so‑ cial classes. Chapter 13 on the CDC (Carvalho, Feferman, Knight and Woroch) describes a common scenario in the country: a regional monopoly in broad‑ band services that resulted in high cost and low quality of services. This was one of the main factors that motivated the investment in the CDC network,

384

flavio feferman

which led to an expansion of services and competition in the state. João Moura (Chapter 12) and Perez and Vale (Chapter 11), argue that the implementation of regulation related to access and competition must be more effective in the country, as entry barriers that favor dominant companies persist, limiting competition and market access by small and mid-sized operators. In Chapter 7, Flavia Lefèvre discusses risks of anti-competitive practices observed in Brazil, such as zero rating, which violates the network neutrality principle, inhibiting competition and innovation. Nathalia Foditsch and Luca Belli (Chapter 5) address the net neutrality theme in the context of wireless broadband, mentioning not only the neutrality of services that utilize the net‑ work, but also “technological neutrality” – a principle that refers to the types of technologies used for electronic communications services. According to the authors, political and regulatory choices should be developed in a way that does not favor specific economic groups. Still on the subject of regulation and competition, Perez and Valley praise the regulation that opened opportunities for small and mid-sized telecom‑ munications companies, transforming Brazil into a unique case with thou‑ sands of regional operators. The case of the Ceará Digital Beltway reinforces this point and demonstrates how public investments in network infrastruc‑ ture can, paradoxically, stimulate competition: the concession process for sharing the CDC infrastructure is opening regional markets in Ceará for these small and medium-sized operators, bringing broadband Internet to the interior of the state. Much like sharing physical infrastructure (via unbundling, public-private partnerships as in the case of the CDC, or other models) encourages com‑ petition within the fixed broadband market, Foditsch and Belli highlight the importance of spectrum sharing for the expansion of mobile broadband. The authors stress the importance of management models that promote the effi‑ cient use and sharing of licensed and unlicensed spectrum: a ‘hybrid’ strat‑ egy for spectrum management [combines] elements of the traditional com‑ mand-and-control model with spectrum market mechanisms (e.g., creation of secondary markets), as well with as ‘commons’ models (e.g., promoting the creation of community networks and the use technologies such as TV White Space (TVWS). This hybrid approach represents an essential step for the de‑ velopment of wireless broadband in the country. Throughout the world, establishing a competitive environment leads to the expansion of access, the improvement of quality and the reduction

conclusions

385

of telecommunication service costs. Moura notes that access regulation in Europe was decisive for the development of alternative operators, generating an expansion of supply, increasing broadband average speeds, and promoting a significant reduction in prices. The opposite happens in countries with lim‑ ited competition: Mexico, a country with a high concentration in the telecom‑ munications industry, has the highest broadband Internet access costs (fixed and mobile) among major Latin American countries (see Chapter 1).

2. INVESTMENTS Brazil faces a chronic problem of low investment, caused by high public spending resulting in persistent budget deficits. Government spending pat‑ terns restrict the resources available for public investments in infrastructure, while financing requirements for the budget deficit raise interest rates, crowd‑ ing out private investment. High interest rates also impact telecommunica‑ tions companies, and especially those that do not have access to subsidized loans from the Brazilian National Economic and Social Development Bank (Banco Nacional de Desenvolvmento Econômico e Social – BNDES). According to World Bank data, between 2000 and 2014, Brazil invested on average 19% of its GDP, well below other large emerging countries such as China (41%) and India (29%).2 It is estimated that during this same period, government investments in infrastructure were on average only 3.5% of GDP, despite ini‑ tiatives such as the Growth Acceleration Program (Programa de Aceleração do Crescimento, PAC), launched in 2007, which increased public investments starting in 2008. The telecommunication sector, despite having evolved significantly over the past two decades, had few new public sector investments to expand the reach, quality and speed of broadband networks, and these only happened after the reactivation of Telebras in 2010. The progress made after the privat‑ ization of the telecommunications sector in 1998 was primarily the result of large investments by private sector companies. As Alimonti points out, the National Broadband Program (Programa Nacional de Banda Larga – PNBL) did not result in the expected public sector investments due to a lack of resources, 2

World Bank, data on gross fixed capital formation as a percentage of GDP. Available at: . Accessed: January 12, 2016. Public infrastructure investments are calculated as total fixed capital formation minus private sector fixed capital formation, also according to World Bank data.

386

flavio feferman

political priority and government implementation capacity. Some of the re‑ sources allocated to the program have not been invested, and the PNBL was also affected by the curtailment (contingenciamento) of budgets by the Ministry of Finance. Tude highlights the importance of the public sector in stimulating invest‑ ments in optical networks (backhaul) for all municipalities, promoting the sharing of dark fibers among service providers and thus enabling mobile and fixed broadband services under a more competitive environment. Coimbra and Silva, as well as Perez and Vale, emphasize the importance of financing programs for small and medium companies, many of these active in smaller markets. The new government program to be launched in 2016, Broadband for All (Banda Larga para Todos), envisages the creation of credit lines by BNDES to finance optical networks for access and transport, encompassing also small and medium providers. Authors such as Alimonti, Barbosa, Lefèvre, Perez and Vale also empha‑ size that infrastructure investments should be complemented by digital edu‑ cation programs to develop competencies and human capital. According to Barbosa, Senne, Oyadomari and Bittencourt (Chapter 6): It is also essential to foster the development of ICT skills needed for effective participation in the digital economy. Thus, greater dissem‑ ination of access to broadband Internet has created a new debate on the socio-economic impacts for connected populations. This debate has increasingly highlighted how individuals appropriate and utilize available network resources.

On this topic, Perez and Vale underscore the contribution of small and medium-sized operators in their role as educators and disseminators of Internet use in the country.

3. TAXATION AND SUBSIDIES Knight, Tude and Moura, among other authors, point to taxation as a signifi‑ cant barrier for the growth of the telecommunications sector in Brazil, as well as for the universalization of broadband Internet. Tude indicates that Brazil ranks among the three countries with the highest tax burden on telecommunications services. He estimates that government taxes and fees account for 30-40% of users’ monthly bills, depending on the amount of value added tax (Imposto

conclusions

387

sobre Circulação de Mercadorias e Prestação de Serviços – ICMS), which varies by state. Moura estimates taxation of up to 50% in the telecommunications sec‑ tor. In a study published at the end of 2015, the Brazilian Telecommunications Association (Associação Brasileira de Telecomunicação – Telebrasil) estimated that in 2016 the average tax burden for telecommunications services would reach 46%, ranging from 40.2% to 63% depending on the state.3 Eleven states and the Federal District increased the ICMS tax rate for telecommunication ser‑ vices in 2016, given the fiscal situation faced by Brazil at the end of 2015. Knight (Chapter 8) notes that raising the ICMS reduces the use of telecommunications services (ceteris paribus) and therefore argues that this tax increase may not gen‑ erate the revenues expected by the states. This tax burden has a direct impact on the profitability and investment capacity of companies in the sector, as well as on the demand for telecommu‑ nications services, since these taxes raise service costs for final users. Moura notes that fiscal and financial incentives have been generously allocated to traditional sectors such as automobiles, meat processing, and appliances. The authors propose a reduction of taxes on the telecommunications sector to stimulate investment, recognizing the important impact of telecommunica‑ tions for economic and social development, as well as for the competitiveness of firms. In addition to the high taxes on the telecommunications industry, Knight notes that the three main funds for the development of the sector (Fust, Fistel and Funttel) have not effectively contributed to the expansion of telecommu‑ nication services. Based on studies by the Federal Senate and the Brazilian Federal Audit Court (Tribunal de Contas da União), Knight argues that the impact of taxation “is even worse because of the diversion of these taxes, by the Ministry of Finance, for other purposes: more than 90% of funds raised from telecommunications companies, by the three sectoral funds created to support the expansion of the sector, are allocated to other ends.” In parallel to this high taxation scenario, Alimonti criticizes the subsidy policies within the telecommunications sector itself, from the point of view of achieving universalization. Even while the sector suffers from low public in‑ vestment, the government offers tax exemptions to large operators. Alimonti argues that

3 Available at: . Accessed: January 15, 2016

388

flavio feferman

If the interest of the federal government [in the PNBL] is to stimu‑ late investment in telecommunications networks, which would pos‑ sibly occur to a lesser extent without the tax exemptions, it must be stressed that public funds are not being collected – funds that could be better directed to the task of universalizing, with quality, the ac‑ cess to broadband Internet in the country, including contributions to Telebras, which is subject to budget curtailments.

Below, we address the issue of universalization.

4. GOVERNMENT ROLE IN UNIVERSALIZATION Several chapters highlight the role of the public sector in the universaliza‑ tion of telecommunications services, through investments, regulation, public concessions, or by acting directly as a provider of telecommunication services (as in the case of Telebras and the Ceará Digital Beltway). Flavia Lefèvre in‑ troduces the concept of broadband Internet as a “public utility” – an essential service that government should guarantee to citizens – and argues that the large private operators do not have the economic incentives to provide uni‑ versal telecommunications services. Similarly, other authors emphasize that private investment alone will not lead to universalization. In Chapter 3, Alimonti argues that the telecommunications service that supports broadband (infrastructure layer) should be provided under both the private and public regimes, which does not occur at present: “Even though broadband is increasingly essential, including for fixed line telephony, its pro‑ vision is carried out exclusively under the private regime, in accord with the interest of the companies, which prefer less regulation”. Alimonti notes that under the private regime the approvals for operation (outorgas) are granted upon authori‑ zation, and the regulation is based on the defense of competition, without obligations related to universalization, reasonable service rates (modicidade tarifária), and revertibility of assets. Currently, fixed telephony is the only telecommunication service provided, in part, under a public regime.

Alimonti proposes that the government participate in the network layer and that “this eminently wholesale service should be operated as a public con‑ cession; the concessionaires would then observe the universalization targets

conclusions

389

related to the deployment of backbones and backhaul.” Similarly, Eduardo Tude argues that the government should establish service goals as part of public tenders, which did not occur, for example, during the 700 MHz fre‑ quency bid carried out by Anatel in 2014, as the government prioritized reve‑ nue collection in this tender. Chapter 13 on the CDC introduces the concept of the “access frontier,” and discusses the role of government and public-private partnerships in the expansion of broadband Internet access in peripheral and rural areas, which have low profitability for large companies (but which constitute attractive markets for small and medium-sized regional operators). Chapter 11, by Perez and Vale, tells the story of these small and medium Internet service providers in Brazil, which were and still are the true “pioneers” of the Internet in the country. These operators are responsible for the deployment of thousands of kilometers of their own fiber optic networks to serve the needs of new region‑ al and local markets. Following the establishment of their association, Abrint, the government has supported these companies, recognizing their import‑ ant role in the universalization of access, whether in the interior of Ceará (as described in Chapter 13 on the CDC) or in Rio de Janeiro’s Rocinha squater settlement (as recounted in Chapter 15 by Coimbra and Silva). In their chapter on the RNP, Grizendi and Stanton describe the role of ac‑ ademic networks in the evolution of broadband Internet in the country. Since the beginning, RNP’s projects have had government support and funding. With this public support, RNP developed several initiatives, including a na‑ tional backbone (Ipê Network), metropolitan area networks in 40 cities across the country (Redecomep), and a new initiative to bring broadband Internet to academic institutions located in the interior regions of states (Veredas Novas). The chapter on the Connected Amazon program (Amazônia Conectada) – by Grizendi, Stanton and Sales – describes the role of different government participants (military, federal government, state agencies), as well as RNP, in a pioneering project involving the deployment underwater cables that will bring broadband Internet to riverside communities in the Amazon region. These cases illustrate the different forms of government involvement for universalization: as a funder of infrastructure or services, as a commercial partner in a public-private partnership, or as a regulator to encourage private sector participation.

390

flavio feferman

5. CIVIL SOCIETY PARTICIPATION A key feature of Internet development in Brazil has been the extensive in‑ volvement of civil society, from the beginnings of the Internet in the 1980s to the recent establishment of the Internet Bill of Rights (Marco Civil da Internet). Demi Getschko (in Chapter 10, on the role of NIC.br) and other authors em‑ phasize the importance of civil society and non-profit institutions in Brazilian Internet governance, including their role in coordination and standardization, which are fundamental for the development of the Internet in the country. This governance structure reinforces the principle of an open and democratic Internet. As previously mentioned, Chapter 9 on the RNP highlights the role of another segment of civil society, the academic community, in expanding broadband access. Civil society continues to play an important role in the movement for universal broadband access, including programs such as “Broadband is your right” (Banda Larga é um direito seu). Alimonti also emphasizes the impor‑ tance of civil society in establishing government policies and in monitoring the implementation of programs: As relevant as the result of the debate on the regulatory model for the sector is its process, which should be carried out openly and with broad participation, given that it relates to a strategic infrastructure and to the future of a service intrinsically linked to the guarantee of fundamental rights in the country.

6. INNOVATION: TECHNOLOGY AND BUSINESS MODELS Innovation appears as a theme in several chapters, either with regard to technological innovation or to business model innovation, including pub‑ lic-private partnerships. Several chapters highlight the role of technological innovation in the evolution of broadband Internet in Brazil. Getschko recounts some of the Internet’s early history in the country and comments on the role of innova‑ tions such as Internet exchange points (Pontos de Troca de Tráfego – PTTs) and their impact on the efficiency of networks. The chapters on the CDC and the RNP demonstrate the impact of fiber optic technologies and dense wavelength division multiplexing (DWDM) systems on the expansion of

conclusions

391

broadband Internet. Grizendi, Stanton and Sales discuss the new generation of international submarine cables with transmission speeds in multiples of 100 Gb/s and DWDM systems. The authors also address the new underwater cable technology for rivers that is being tested and deployed in the Connected Amazon program – a technology that could be utilized in other areas with difficult overland access. All these cases demonstrate the importance of knowhow and investments in new technologies to keep pace with the global broad‑ band growth curve and to promote universal access. Several chapters address innovation in business models. For example, dif‑ ferent authors emphasize the importance of the infrastructure sharing model to achieve economies of scale in network investments and to foster competi‑ tion in markets. The chapters on the RNP and the CDC also illustrate inno‑ vative Brazilian cases involving partnerships among government, businesses and academic institutions. The CDC’s business model serves as an example for many other states that are planning or implementing their own fiber optic networks. The chapter on the CDC also describes the creativity and innova‑ tion of small and medium-sized regional operators, who are extending fiber optic networks to new areas and developing profitable business models for these markets. Still on the topic of innovation, other authors emphasize the importance of competition for innovation, a concept developed over generations by academic scholars such as Joseph Schumpeter, Michael Porter and Clayton Christensen, among others. Competition promotes new ideas, new products and services, not only from new entrants but also from dominant firms, which are encour‑ aged to evolve and create new forms of value for their customers. Some au‑ thors mention the “disruptive” innovations of over-the-top (OTT) companies, which utilize broadband Internet to offer multimedia services such as IP tele‑ phony and video conferencing (companies like Skype/Microsoft, Apple, and WhatsApp/Facebook), streaming video and IPTV (YouTube/Google, Netflix, Amazon), and social networks (Facebook, LinkedIn), among others. Some of these services compete directly with the traditional business models of the telecommunications industry, but on the other hand generate massive de‑ mand for data transmission, which benefits operators. The convergence between telecommunications and content has led to a battle between the telecom operators and OTT companies in Brazil. The OTT companies not only provide services such as IP telephony, which com‑ pete directly with the operators’ traditional services, but are also beginning

392

flavio feferman

to enter the market for the provision of access, which presents a new threat to traditional companies. Examples include recent initiatives by Facebook and Google. This is a pending issue that should be monitored by regulators. (As Lefèvre points out, there are also net neutrality issues, such as initiatives that utilize the zero rating model.) On the other hand, this new market dy‑ namic tends to increase competition and innovation, benefiting consumers and expanding access. These OTT services based on new content and business models represent one of the major vectors of growth and innovation in the modern economy, driven not only by giant companies like Facebook and Google, but also by thousands of startups that offer innovative services and content via broadband Internet. Many of these companies are transforming markets and democra‑ tizing access to services, including services with social impact such as Khan Academy (low-cost distance education), M-Pesa (universal access to financial services via mobile) and Skype (low cost telephony), among others. Brazil has several entrepreneurial hubs focused on information technology, such as Porto Digital (Recife), Belo Horizonte, Florianópolis, São Paulo and Rio de Janeiro. Access to quality broadband is an essential factor for the competitive‑ ness of these technological hubs and for the creation and growth of innovative technology companies in Brazil.

FINAL CONSIDERATIONS: STRATEGIC VISION Despite the initiatives and advancements discussed throughout the book, major challenges remain for the expansion and universalization of broadband in Brazil. As Alimonti points out, the PNBL failed to achieve the expected objectives, and in early 2016 the government had not yet defined clear pa‑ rameters and goals for the new Broadband for All program.4 More generally, Knight (Chapter 1) argues that Brazil needs a more holistic strategy of infor‑ mation technology for economic development (ICT4D), with specific objec‑ tives, resources, effective implementation mechanisms, as well as systems for monitoring progress and evaluating impact. This strategy should address the issues discussed previously: regulation for competition and innovation, investment, tax reform, universalization (including both access and capacity 4 Available at: . Accessed: January 15, 2016.

conclusions

393

building for technology use), as well as innovation and the participation of civil society. The importance of a coherent broadband strategy for the country should not be underestimated. Several studies cited throughout this book indicate a direct relationship between access to broadband, economic growth and so‑ cial inclusion. For example, a study on Brazil indicates that broadband ex‑ pansion is associated with an increase of 1,0-1.4% in employment growth rate. Another study, by the World Bank, indicates that among developing countries, an increase of 10% in broadband penetration adds 1.38% to GDP growth. Therefore, the expansion of broadband is precisely the type of tech‑ nology strategy that promotes economic convergence between rich countries and developing nations. Every sector of society has a role in this transformation: government, busi‑ ness, academia, industry associations and civil society. Similarly, we imag‑ ine that each reader of this book recognizes the importance of this issue and could, in some way, contribute to the debate and to the evolution of broadband in Brazil.

about the authors

ABRAÃO BALBINO E SILVA is Anatel’s manager for monitoring of relations among providers. He is also rapporteur for the study group on competition policy, relevant markets and market power of the International Telecommunication Union (ITU). In 2011 he won the SEAE Award for Economic Regulation, with the text “Remuneration of Mobile Networks in Brazil.” At Anatel he is responsible for the preparation, deployment and updating the General Plan on Competition Targets (Plano Geral de Metas de Competição – PGMC). He is a Communication Networks engineer and holds master in Regulation and Business Management from the University of Brasilia (UNB). [email protected] ALEXANDRE BARBOSA is manager of the Regional Centre of Studies for the Development of the Information Society under the auspices of UNESCO (Cetic.br). He holds a degree in Electrical Engineering from the Catholic University of Minas Gerais (PUC-MG), a Master’s degree in Computer Science from the Federal University of Minas Gerais (UFMG), a MBA from the University of Bradford, a PhD in Business Administration from the Getulio Vargas Foundation, and did Postdoctoral work in Business Administration at HEC Montreal. [email protected] ALISSON BITTENCOURT is a research analyst at the Regional Centre of Studies for the Development of the Information Society (Cetic.br), linked to the Brazilian Internet Steering Committee (CGI.br). He is mainly active in the ICT Households and ICT Health projects. He has a Bachelor’s degree in Social Sciences from the University of São Paulo (USP). [email protected]

396

BROADBAND IN BRAZIL

ASAFE COIMBRA is technical director and co-owner of Net Rocinha, an Internet service provider in one of the best-known favelas in Rio de Janeiro, where he was born and lives. He finished high school and then worked in several companies as a delivery boy, learning to repair appliances and electronics devices. He was introduced to computers when working at a company in the sector. In February 2006, he decided to set up his own business, starting with a lan house. He migrated to the computer business, repairing and selling computers and accessories. In 2010, he became co-owner of Net Rocinha with Samuel Pereira. [email protected]

BASILIO R. PEREZ is President of the Advisory Council of the Brazilian Association of Internet and Telecommunications Providers (Abrint), an access pro‑ vider since 1995, Chief Operating Officer of BCMG Internet Ltda partner and owner of Carrier Micro Group Ltda. He holds an MBA in Strategic Management from the University of São Paulo (USP) and is an electronic engineer from the University Santa Cecilia (Unisanta). [email protected]

BRENO VALE is executive director at PLUG Telecom and Executive Vice President of the Brazilian Internet and Telecommunications Providers Association (Abrint). Bachelor in Business Administration from the University of Minas Gerais (UEMG) and techni‑ cian in electronics from the Technological Orientation Educational Center of West Minas (CoteomMG). bre‑ [email protected]

about the authors

397

DECÍLIO SALES is a general in the Brazilian Army, Chief of the Army’s Integrated Telematics Center (CITEX) of the Army’s Department of Science and Technology and director of the Connected Amazon program. He was a professor at the University Severino Sombra (USS) and of the Military Engineering Institute (IME). He holds doctorate degrees in Electrical Engineering from the Catholic University of Rio de Janeiro (PUC-Rio) and Naval Sciences from the Naval War College. decilio@ citex.eb.mil.br DEMI GETSCHKO is CEO of the Information and Coordination Nucleus of Dot BR (NIC.br), member of the Brazilian Internet Steering Committee (CGI.br) and associate professor at the Pontifical Catholic University of São Paulo (PUC-SP). He worked at the Centre for Electronic Computing at the University of São Paulo (USP) and the Foundation for the Support of Research of São Paulo State (FAPESP) and is considered one of the fathers of the Brazilian Internet. He was member of the board of the Internet Corporation for Assigned Names and Numbers (ICANN). He is an electrical engineer from PoliUSP, and holds master and PhD degrees in Engineering from the same institution. [email protected] EDUARDO GRIZENDI is Director of Engineering and Operations of the National Education and Research Network (RNP) and professor at the National Telecommunications Institute (Inatel). He worked for companies and institutions in the area of telecommu‑ nications and innovation, including CPqD and Agencia Inova Unicamp. He is an electronic engineer from the Technological Institute for Aeronautics (ITA), holds a master’s degree in telecommunications from Inatel and an MBA from the Getulio Vargas Foundation (FGV). He is the author of a manual for Brazilian ICT companies, promoted by Softex and published by Editora Publit, Rio de Janeiro. [email protected]

398

BROADBAND IN BRAZIL

EDUARDO TUDE is President and founding part‑ ner of Teleco, a telecommunications consulting firm, an analyst of the Brazilian telecom market, he is a ju‑ ror for the Global Mobile Awards and acted as Expert Visiting Professor at the State University of Campinas (Unicamp) in 2013. A pioneer in the development of satellites in Brazil at Inpe, he occupied various corpo‑ rate positions in companies such as: VP of operations at BMT, chief operating officer of Pegasus Telecom and mobile planning man‑ ager at Ericsson. He is a telecommunications engineer (IME) and Master in Telecommunications (INPE), having worked in the areas of optical networks, mobile systems and satellite communications. [email protected] FÁBIO SENNE is Research Coordinator at the Regional Centre of Studies for the Development of the Information Society (Cetic.br), linked to the Brazilian Internet Steering Committee (CGI.br). He was one of the technicians responsible for thematic media analyses conducted by the News Agency for Children’s Rights (ANDI). He collaborated with the Center for Studies on Media and Policy (NEMP) at the University of Brasilia. He holds a Master’s degree in Communication from the University of Brasilia (UNB) and a BA in Social Sciences from the University of São Paulo (USP). [email protected] FERNANDO CARVALHO is a professor of Computer Sciences at the Federal University of Ceará (UFCE). He was president of the Information Technology Company of Ceará (Etice) for eight years, where he coordinated the design and implementation of the Ceará Digital Beltway (Cinturão Digital do Ceará). He participated as coordina‑ tor of the consortium for software applications linked to the Ministry of Communications for developing the Brazilian Digital Television System (SBTVD). In 1992 he obtained a PhD from the University of Montpellier, and in 1989, an MSc by the Federal University of Paraíba (UFPB), both in Computer Sciences. [email protected]

about the authors

399

FLÁVIA LEFÈVRE is a partner of the Lescher and Lefèvre Advogados Associados Law Firm, counselor and attorney for Protest – Consumers Association. She represents civil society in the Brazilian Internet Steering Committee (2014-2017) and was representa‑ tive for users on the Advisory Council of Anatel (20062009). She has been a member of the Directorate for Telecommunications Infrastructure of the São Paulo Federation of Industries (FIESP) since 2011. She holds a Master’s degree in Civil Law from the Pontifical Catholic University of São Paulo (PUC-SP). fla‑ [email protected] FLAVIO FEFERMAN has been a professor at the University of California, Berkeley, Haas School of Business, since 2008. In 2014 he received the Earl F. Cheit award for excellence in teaching. He is also presi‑ dent of Developing Markets Group, which specializes in consulting projects in the US, Latin America and Brazil. During his 25 years of professional experience, he direct‑ ed several international projects focused on economic development, technology and business strategy for organizations such as the World Bank, IADB, and USTDA, as well as for corporations and NGOs. He is a graduate of Stanford University and U.C. Berkeley, with advanced degrees in Economics and Business Administration. [email protected] GLENN WOROCH is a Professor of Economics at the University of California, Berkeley, and Executive Director of the Center for Research on Telecommunications Policy at Berkeley. He was Professor of Economics at the University of Rochester and Economist of GTE (Verizon) Laboratories. He currently participates in the editorial committee of the Journal of Telecommunications Policy and serves as Senior Policy Scholar at the Center for Business and Public Policy of Georgetown University. Professor Woroch pub‑ lished several articles on telecommunications policies and acts as a consultant for companies in the sector. He holds PhD and MA degrees from the University of California, Berkeley. [email protected]

400

BROADBAND IN BRAZIL

JOÃO MOURA is the CEO of Telcomp, an association that brings together 55 telecommunications operators in Brazil, including all the integrated large ones and more special‑ ized operators in specific market segments. He was a part‑ ner of Coopers & Lybrand (now PricewaterhouseCoopers), specializing in corporate strategy and finance, and held excutive positions in such companies as BellSouth-BCP Telecommunications, Paramount and Terphane. He is an economist holding an MBA from Coppead Institute of the Federal University of Rio de Janeiro (UFRJ) and has completed executive courses at the Sloan School of Management of the Massachusetts Institute of Technology (MIT) and Stanford University. [email protected] LUCA BELLI is a senior researcher specializing in Internet governance at the Center for Technology and Society (CTS) at the Law School at the Getúlio Vargas Foundation in Rio de Janeiro (FGV Direito Rio) since 2015. He worked as a specialist at the Internet Governance Unit of the European Council and has consulted for the Internet Society (ISOC). He is also a founder and co-director of the Dynamic Coalition on Internet Neutrality of the Internet Governance Forum (IGF) and is one of the editors and senior authors at Medialaws.eu. Belli holds a PhD in Public Law from the Université PanthéonAssas (Paris II). [email protected] MICHAEL STANTON has been Director for Research and Development of the National Education and Research Network (RNP) since 2002. He taught in the areas of Computers and Communication at the Federal Fluminense University (UFF) between 1994 and 2014, of the Pontifical Catholic University of Rio de Janeiro (PUCRio) between 1973 and 1999, and the Technological Institute of Aeronautics (ITA) between 1971 and 1972. He obtained his bachelor and doctors degree in mathematics from Cambridge University between 1964 and 1971. [email protected]

about the authors

401

NATHALIA FODITSCH (LLM/MPP) is an attorney specializing in politics and communications regula‑ tions. She was a researcher at the Center for Technology and Society (CTS) at FGV Direito Rio, a consultant for the Aspen Institute Communications Policy Program and the Special Program for Broadband of the InterAmerican Development Bank (IDB), having also worked in the Administrative Council of Economic Defense (Cade) in Brazil. She is an Invited Professor of the course “Comparative US Brazil Legal and Judicial Systems” at the Washington College of Law. She is currently a doctoral student in Internet Governance at American University in Washington DC, sponsored by a grant from Google. [email protected] PETER KNIGHT is an economist specializing in the use of Information and Communication Technologies to accelerate social and economic development. He is a founding member, researcher and Board member of the Fernand Braudel Institute for World Economics in Sao Paulo and author and/or organizer of seven books on the Internet and development. He has held a variety of technical and management positions at the World Bank, Cornell University, Ford Foundation, Brookings Institution and at the Training Center for Economic Development (CENDEC) in Rio de Janeiro. He has a PhD from Stanford University and graduated from Oxford University and Dartmouth College. [email protected] SAMUEL SILVA is co-owner of Net Rocinha, an Internet service provider in Rocinha, one of the best known favelas in Rio de Janeiro, where he was born and lives. In 2007, he developed a large clientele act‑ ing as informal Internet services provider in that com‑ munity. In 2011, with Asafe Coimbra, founded Net Rocinha, a company that holds a license for Multimedia Communication Services (SCM) from Anatel. In June 2015, looking for more interaction with the business community, he ran for and was elected a member of the Advisory Board of the Association of Internet and Telecommunications Providers (Abrint). [email protected]

402

BROADBAND IN BRAZIL

VERIDIANA ALIMONTI is a member the Board of Directors of Intervozes – Brazil Collective for Social Communication. She was a representative for civil soci‑ ety in the Brazilian Internet Steering Committee (CGI. br) between 2011 and 2013, consumer representative on the Defense Committee of Telecommunications Service Users (CDUST) at Anatel until early 2015 and attorney for the Brazilian Institute for Consumer Defense (Idec) between 2010 and early 2015. She is a lawyer and holds a Master’s degree in Economic Law from USP with a study on public communication policies in Brazil. [email protected] WINSTON OYADOMARI is Research Analyst at the Regional Centre of Studies for the Development of the Information Society (Cetic.br), linked to the Brazilian Internet Steering Committee (CGI.br). He worked on sev‑ eral Cetic.br research projects and since 2013 has been co‑ ordinating the ICT Households research. He holds a de‑ gree in Public Administration from the School of Business Administration of the Getúlio Vargas Foundation of São Paulo (FGV-EAESP). [email protected]

font:

Scala

#Figurati in social media