Oct 28, 2004 - a vision of future 4G wireless systems, arguing that such systems will rely heav- ily on the packet-switched IP platform to unify heterogeneous ...
LYT/BOOKS/NOV
10/28/04
8:56 AM
Page 14
BOOK REVIEWS (Continued from page 12) Chapters 3 and 4 address RRM issues for 2.5G and 3G systems. 3G networks aim to serve disparate traffic types, which is very different from 2G cellular systems that are primarily voice-oriented. As such, congestion problems in 3G networks are more complex and tend to occur more frequently. This chapter attempts to cover these problems in perspective, including the topic of congestion control for 3G wireless networks. Perhaps the chapter of interest to many researchers is Chapter 5, which addresses RRM for systems beyond 3G. The chapter starts out with a useful comparison of 1G, 2G, 2.5G, 3G, and 4G systems. The authors then provided a vision of future 4G wireless systems, arguing that such systems will rely heavily on the packet-switched IP platform to unify heterogeneous wireless systems. Adaptivity and software programmability are expected to be key features of 4G systems. A global RRM solution is therefore needed to improve network performance by decreasing the effect of traffic overload. However, it would have been more useful if the authors included RRM for emerging
14
wireless technologies such as ultrawideband and multiple-input multipleoutput antenna systems. RRM for such systems can be particularly challenging. It is also unclear to the reader the role of RRM for software radio systems since radio spectrum is becoming increasingly deregulated, potentially leading to an abundance of bandwidth when spectrum is used (and reused) more efficiently and cooperatively by software radio systems. The final chapter (Chapter 6) emphasizes the role of economics on RRM from both the subscriber and network operator perspectives. It advocates the use of a business model before the execution of an RRM technique. However, the trade-off between user satisfaction and RRM effectiveness should be considered carefully in the design of a scalable business model. Overall, the book covers a broad spectrum of traditional, contemporary, and emerging RRM methods, and should appeal to a wide range of readers interested in the subject. It might not be an ideal textbook because it does not include problem sets or exercises, but the practical approach can easily complement other more analytical texts
on RRM. Rather than just stating the facts, the authors explain things and attempt to integrate the use of RRM in disparate wireless technologies. It is timely, well written, and well organized, and should be a useful reference for any wireless practitioner or researcher. POLICY-BASED NETWORK MANAGEMENT, SOLUTION FOR THE NEXT GENERATION BY JOHN STRASSNER, MORGAN AND KAUFMANN, 2004, HARDCOVER, 516 PAGES, ISBN: 1-55860-859-1 REVIEWER: JACEK WSZOLEK Policy-based management (PBM) is the subject of fast development in universities, standardization bodies, and within industry. Currently we can see a trend to integrate many (or even all) network services in a single large network. This approach increases network complexity and results in a complicated and painful network configuration process. It is expected that policy-based network management (PBNM) will be an answer to this challenge. Strassners’s book explains exactly what PBNM is, summarizes what has been done in this subject area up to now, and proposes new directions in PBM. The solutions described in this book are mostly based on work in the TeleManagement Forum, but it also presents work done by the Internet Engineering Task Force (IETF) and Distributed Management Task Force (DMTF). It is written in a way that is accessible to a wide readership, but mainly it is targeted to developers of new network architectures and people who want to understand what PBM is. Chapter 1 contains an introduction to the PBM concept. It briefly defines PBNM by first explaining what it is, why there is a need to change the way network management is done, and what kind of benefits it will provide. The second chapter provides a foundation for understanding PBM. By presenting the conceptual policy model, the author familiarizes the reader with the fundamental building blocks of a PBM system, which allows him to introduce and define policy terminology. Next, the requirements of a PBNM system are presented. The main emphasis is placed on integrating the business and networking worlds. The goal of this task is to enable business requirements to drive the configuration and management of network services. This is done by a set of mappings between policies at different abstraction levels (from the business to the network and even the instance level). This set of mappings (Continued on page 16)
IEEE Communications Magazine • November 2004
LYT/BOOKS/NOV
10/28/04
8:56 AM
Page 16
BOOK REVIEWS (Continued from page 14) forms the policy continuum (also defined in this chapter). Chapter 3 concentrates on information in the PBNM system. In the beginning an information model is presented. It is used to provide common representation of information in an object-oriented way. It is considered to represent something more than just data. In the information model, apart from expressing objects and their semantics, interactions between them are also modeled. Because the information model is not sufficient to ensure sharing and reusing of common data, the concepts of languages and a data dictionary are presented. Languages are responsible for translation between different levels of the policy continuum. To do this a data dictionary is used. After explaining the information subject matter, the author presents existing approaches to sharing information (DEN, IETF, and DMTF approaches), summarizes them, and recommends a new approach, DENng, which is being standardized by The TeleManagement Forum. In Chapter 4 the author discusses how policies are applied and operate in the PBNM system. After presenting the communication, categorization, and assignment issues of policies, policy operation is overviewed in detail. Two major concepts are described: physical and logical system requirements, and the effects of applying policy with special emphasis on conflict detection and resolution. Chapter 5 covers detailed requirements for using high-level policies, and
16
describes their structure and functions in a PBNM system. It introduces the Ponder system and its four basic types of policies: authorization, refrain, obligation, and delegation. Using Ponder as a background, the author presents DENng enhancements to this concept. Special emphasis is placed on events, constraints, and roles in DEN-ng. In the final section of Chapter 5 the use of a high-level policy is presented on the basis of an multiprotocol label switching (MPLS) virtual private network (VPN) example. Chapter 6 describes the motivation for creating a new policy model, which is then described in Chapter 7. The author presents an implementation of the policy concept in the DEN-ng model. Several important aspects of the approach used in this model are described. First, grouping policies using PolicyGroup and PolicySet classes is presented. The PolicyRule class, representing policy rule, is described in detail. Further, a PolicyStatement class that models the triplet variable, operator, and value is characterized. This class is fundamental for the creation of policy conditions and actions. Additionally, the way of modeling the concepts of policy event, policy condition, and policy action in the DEN-ng model is presented. At the end of this chapter, interactions of the policy model with other DEN-ng models are analyzed. Chapter 8 concentrates on showing the PBNM system architecture. The author takes several example policies and presents their complete life cycle from installation to execution through
the system. It facilitates understanding the requirements for the PBNM system. It also shows that the simplistic IETF model lacks additional components and requires architectural changes. A new architecture is then presented, and each component, such as the policy repository, policy broker, and policy server, consisting of a set of policy decision points and policy enforcement points, is discussed in detail. Special emphasis is placed on pointing out how the use of three building blocks of the PBNM system, an information model, a policy language, and a data dictionary, binds the theory to implementation. Chapter 9 contains an example of how using the DEN-ng information model enables various levels of policy continuum development, and how those levels complement one another and work together to implement policies. The author shows it by using an example of an MPLS VPN. As an introduction, a detailed description of the VPN concept and a brief guide to the DEN-ng model are presented. It makes the reader familiar with the service and the tools used to implement it. Chapter 9 presents modeling an MPLS VPN in DEN-ng at different levels of the policy continuum, from the business through the system to the implementation view. The next two chapters (10 and 11) present several examples of using a policy. By using these examples, Chapter 10 explains that policy is something more than just changing commands in configuration files and shows a broad outline of its full potential. The last chapter of the book (Chapter 11) summarizes what has been done in PBM and presents the current state of the art. It introduces companies working on PBNM solutions, like Gartner and Burton Groups and standardization bodies (especially TMF), and presents three vendor commercially implemented examples: Nortel’s Optivity Policy Services, Anaya’s Policy Manager (CajunRules), and Cisco’s QoS Policy Manager from CiscoWorks network management software. This book is a unique and very good publication on policy-based network management. It is successful in explaining what PBNM is and what it offers. PBNM is a very wide subject and requires a considerable amount of additional knowledge to have a good grasp of it. The author attempts to address all these requirements and explain them. This enables good understanding of this topic, although it makes the book hard to read. It is a useful and obligatory reference for engineers, researchers, and students interested in PBNM.
IEEE Communications Magazine • November 2004
