Sustaining Life During the Early Stages of Disaster ... - IEEE Xplore

SAKURAI_LAYOUT.qxp_Layout 1/6/14 2:29 PM Page 176

ITU KALEIDOSCOPE

Sustaining Life During the Early Stages of Disaster Relief with a Frugal Information System: Learning from the Great East Japan Earthquake Mihoko Sakurai, Keio University Richard T. Watson, University of Georgia Chon Abraham, The College of William and Mary Jiro Kokuryo, Keio University

ABSTRACT Important lessons for responding to a largescale disaster can be gleaned from the March 11, 2011 Great East Japan Earthquake and tsunami. The failure of the electrical power system and the resultant loss of information communication and processing capability severely constrained the recovery work of many municipalities. It was difficult for supporting organizations to collect and share information. A frugal information system designed around the four u-constructs is suggested as a solution for handling the very early stages of disaster relief, typically within the first 72 hours and even upon the realization of an impending disaster. This article focuses on basing communications on the most frequently available device, the cellular phone, as the foundation for a frugal IS for disaster relief. Familiar and available tools place minimal stress on an already strained communication system, and enable effective connection between those impacted by a disaster and those involved in disaster relief.

AN INFORMATION SYSTEM FOR DISASTERS March 11, 2011 (3.11) marked a day of devastation for East Japan. An earthquake and tsunami immobilized much of the technologically advanced nation. Nature managed to subdue the prowess of some of humans’ greatest inventions over the course of the disaster. In the wake of a catastrophe, the first few hours and days are critical, but typically thwarted by communication and power interruptions that threaten the sustainment and sanctity of life. The most important lesson from the Great East Japan Earthquake is that it is impossible to build a system that never fails. Rather, we need

176

0163-6804/14/$25.00 © 2014 IEEE

to create resilient information systems (ISs) that can quickly regain functionality to perform critical post-disaster missions and return smoothly to full processing capability thereafter. We should incorporate this objective into requirements, especially for ISs serving disaster prone regions. The lack of communication capabilities and information processing resources needs to be swiftly handled by deploying a system with basic capabilities and sufficient bandwidth to meet the most pressing needs while leveraging a basic communication tool most people possess. A frugal IS [1] embodies a set of characteristics that enables swift and effective deployment of a very limited IS designed, in this case, to gather and distribute minimal but critical information to maximize survival rates. Most articles on disaster relief discuss actions after the first 72 hours (three days), when there has been time to mobilize resources from outside the affected area [2, 3]. These studies also focus almost solely on the information and communication technologies (ICT) division as the primary user, systems architecture, and technical components to retain power and communication at the cell site. This differs in that it not only addresses the aforementioned issues but also includes emphasis on the immediate basic communication capabilities and information needs of the individuals directly impacted by the disaster, as well as those involved in disaster relief. This work focuses on how municipal governments can address the situation by themselves immediately after a major disaster. This article is also unique in that we place an emphasis on the connectivity at a higher layer (i.e., data integration) in addition to the lower layer connectivity that most other articles address. We know of no other work that deals with municipal government data management immediately after the March 11 Great East Japan Earthquake.

IEEE Communications Magazine • January 2014



Fukushima Prefecture

The Great East Japan Earthquake occurred at 14:46 Japan Standard Time on March 11, 2011. At a Richter scale of 9.0, it was the largest earthquake on record for Japan. More damaging than the quake itself, a tsunami up to 40 m high hit the coastline, devastating cities and towns. The Fire and Disaster Management Agency reported 16,131 deaths, 5994 injuries, and 3240 missing as of January 2012. It also reported 128,497 houses totally lost and more than 900,000 partially destroyed. From an ICT perspective, disruption of communication and loss of IS capabilities for operations were a significant hindrance to effective and rapid recovery. People and organizations were deprived of the information and processing systems required to deal with the situation. The effect was particularly noticeable at the municipal government level, because it is its responsibility to support its citizens in an emergency. Loss of power (Table 1), termination of telecommunications services (primarily due to the loss of power to cell sites), and data loss were the key problems. The extent of the damage far exceeded any predictions, and recovery efforts had to be made outside of prepared procedures. Rescue workers had to rely on their judgment and the capabilities of personnel on the scene. Loss of power and communication especially affected all initial relief efforts. Some local governments were equipped with emergency power generators, but these covered only basic needs and were not sufficient to enable ICT to function at the needed level. Recovery of commercial power required more than four months in some areas. Telecommunications were also disrupted as a result of the power outage. Many switching facilities were lost and cables were damaged. Recovery time varied depending on the damage. Recovery of fixed line telephony and Internet reconnection required from one week to four months (Table 2). The Government Disaster Management Radio Communication Network and the satellite phone system survived. Municipal governments

Miyagi Prefecture

IMPACT OF THE DISASTER

Municipalities surveyed

Iwate Prefecture

After reviewing a comprehensive report on the impact of the earthquake on ICT at the municipal government level [4], we propose a design for a cellphone-based frugal IS to be deployed during the first phase of disaster relief. The article is structured as follows: • An overview within 13 municipalities of the crisis and its impact on ICT and emergency response capability • Analysis of the implications and a broad statement of requirements for a better solution • Conceptualization of a frugal IS and presentation of the four information drives to describe the design of an IS for first phase emergency response • Matching the design against the specific needs of this crisis identified in the prior section • Technical options • Conclusion

Power resumption timing (days after the disaster)

Miyako City

15 days

Rikuzentakata City

3 days (only areas where emergency response headquarters were set up)

Kamaishi City

Approx. 120 days (March 20 to server room and peripherals)

Otsuchi Town

Approx. 20 days (to the Central Community Hall), approx. 45 days (to the temporary office)

Sendai City

1 day (2 days to the Information Systems Center)

Ishinomaki City

15 days

Kesennuma City

6 days

Higashimatsushima City

4 days

Minamisanriku Town

Approx. 80 days (temporary office)

Iwaki City

No power loss

Minamisoma City

No power loss

Futaba Town

No power loss

Namie Town

1 day

Table 1. Timing of power supply resumption (at the municipal government office).

used these to request support and to organize relief. Their usefulness was, however, limited in that the former connected few destinations and use of the latter was limited due to the high cost. The lack of daily use of satellite phones meant that many did not contain current contact lists. As the number of satellite phones was less than a handful in each location, their use was limited to high-priority external communications. A local government WAN, the usual communication system, was not restored for at least two months in the 13 municipal governments surveyed, and reestablishment took more than a year in the most severely hit municipality. Among individuals, the cellular phone was the most widely used communication tool. The service was available in most areas until the batteries at the cell sites discharged. Conversation was mostly impossible, but packetized email systems could be used immediately after the quake. Many municipal government officials had learned about the coming of the tsunami with TV tuners on their private phones. Cell site batteries ran out by the following day (March 12). They were restored quicker than fixed line communication lines, but nevertheless cell sites were out of service for a couple of days to approximately two weeks. In some areas, mobile cellular phone cell sites were sent 72 hours after the disaster.

177


In summary, ICT infrastructure destruction severely limited communications bandwidth. Under such circumstances, the primary operations of municipal governments following the disaster were: • Confirming the whereabouts and safety of residents • Establishing and operating evacuation centers • Transporting and managing relief goods • Supporting evacuees and creating evacuee lists • Issuing disaster victim certificates These tasks are different from the daily operations of municipal government, and other tasks were suspended to meet the impending and mounting needs of citizens. This places a particular focus on the situation Status of usage *1 (March 11) Municipalities surveyed

immediately following a disaster and before the arrival of outside relief. In such times, the locals have to deal with the situation themselves by employing whatever resources are available to them. What does it take to improve their capabilities to help themselves under extreme duress and in the immediate wake of a disaster? The following sections address this question.

ANALYSIS OF THE FINDINGS Three basic functions are necessary in order to execute the five primary operations identified in the previous section: • Identification of individuals • Compilation of a disaster victim database • Matching of relief goods demand and supply (including usability of roads)

Timing of restoration (Days after the disaster)

Use of satellite mobile phones*3

Iwate Prefecture Miyagi Prefecture Fukushima Prefecture

Landlines

Mobile phones*2

The Internet

Miyako City

×

D (1)

×

Approx. 20 days

—

15 days



Rikuzentakata City

×

×

×

Details unknown

7 days

120 days at the temporary office



Kamaishi City

×

×

×

7 days

7 days

9 days



Otsuchi Town

×

×

×

Approx. 45 days

9 days

Approx. 70 days



Sendai City



D (2)

×

—

3 days

2 days



Ishinomaki City

×

×

×

15 days

15 days

15 days



Kesennuma City

×

D (3)

×

10 days

Approx. 10 days

6 days



Higashimatsushima City

×

×

×

6 days

Approx. 20 days

6 days



Minamisanriku Town

×

×

×

Approx. 20 days

Approx. 20 days

Approx. 20 days

No arrangements

Iwaki City





×

—

—

1 day



Minamisoma City

 (From March 12 ×)

8 days

8 days

8 days

Poor connectivity

Futaba Town



D (4)

×

—

7 days

2 hours

No arrangements

Namie Town

×

×

×

Details unknown

Details unknown

80 days



Landlines

Mobile phones

The Internet

Could not be used: ×, Could be used: , Could be used with some restrictions: D Information on the status of usage of mobile phones is as stated by the survey respondents. The status of usage of mobile phones immediately after the disaster and the timing of restoration varies by telecommunications service provider and area. *3: The user was not always the ICT division. (1) Mobile phones could be used between only a few telecommunications service providers. (2) Varies by telecommunications service provider and area. (3) Could be used until around 10 p.m. on March 11. (4) Could be used only to send and receive emails, not to make phone calls *1: *2:

Table 2. Status of communication means and timing of restoration.

178



In the initial phase, each evacuation center had to record its evacuees. In reality, identification was done by self-declaration. Japanese governmental services use four basic elements of information (name, gender, birth date, and residential address) for individual identification. There is not a unique number for each person. Thus, when evacuation centers received an evacuee, his/her name, address, and birth date were recorded using pencil and paper. In the absence of power, handmade posters were used to publicly list the survivors at a center. Each city had up to several hundred evacuation centers, and fragmentation of data records became an issue. Although there was a strong need for sharing the manually created data lists, this was very difficult because some of the necessary infrastructure was obsolete years ago (e.g., carbon paper). This lack of a single view of the data was particularly problematic as many of the evacuees moved in search of better conditions and food. Information sharing between municipal governments was another major bottleneck. The lack of residential record data hindered police and the self-defense force’s life saving operations. Similar communication failures existed among logistics organizations. The inland city of Tono, which played a major role as a neighboring city to many of the most seriously affected coastal cities, suffered from inaccurate or outdated information on the supply needs of its neighbors. The temporal integration of the victim database was another big issue. This database is the basis for issuing disaster-victim certificates. Victims need to retain these certificates for up to two decades to receive government support. Initially, there was simply a need to record life or death information together with a residential address. Later, information such as damage to a person’s home had to be added. This database had to be maintained as evacuees relocated, and various parties in dispersed locations amended it. To enable delivery of relief, evacuee lists should include a person’s residential status to check eligibility. The fraudulent receipt of relief was a sad reality that had to be addressed. While tentative relief was provided even for those who could not be verified, other more costly operations such as the allocation of temporary housing and loans for housing reconstruction required eligibility confirmation. The baseline data were the residential records maintained by municipal governments. While the local government was responsible for issuing relief certificates, it was not a critical first response issue, but something to handle later. In spite of the importance of maintaining an integrated database, there was fragmentation and duplication at many locations, which often resulted in the establishment of ad hoc databases for each task. In addition to the power and communication loss, stringent security hindered the use of official residential records. Vast amounts of donated goods were delivered to disaster areas, but it was very difficult to determine what was needed where. Many items


were unused and piled up in school gymnasiums. One of the issues was the lack of information regarding the usability of roads. The tsunami and earthquake made many of them unusable for cargo transport. Identification of individuals and matching of relief goods supply and demand can be abstracted as: • Application of an ID to objects, including persons and goods • Recording of time, location, and status of each object with an ID Also, it is highly desirable (if not always possible) that such records be recorded in a single database.

While tentative relief was provided even for those who could not be verified, other more costly operations such as the allocation of temporary housing and loans for housing reconstruction required eligibility

A DESIGN FOR A FRUGAL IS FOR THE FIRST PHASE OF EMERGENCY RELIEF

confirmation.

As the prior analysis shows, immediately following a disaster there is often a lack of communication capabilities and information processing resources. We must rethink how we can prepare ICT to deal with future disaster situations. It is not realistic to assume that each municipal government can prepare for all risks independently. A resilient information system is needed. This situation might be best handled by quickly deploying a frugal IS, which is defined as “…an information system that is developed and deployed with minimal resources to meet the preeminent goal of the client” [1]. Because of the likelihood of limited communication and processing resources, the IS should use minimal bandwidth to send and receive a few critical messages. Furthermore, the frugal IS should focus on the dominant problems of the first few days of a disaster: determining the location and condition of the victims, helping them get emergency support, and identifying the missing. Such information is essential to managing the relief effort and informing concerned relatives and friends.

STRUCTURAL FACTORS Many parties have a potential role in emergency relief, from neighbors to international agencies such as the Red Cross. The major players in most cases are the national government and local governments, such as municipalities, because they have prime but differing responsibilities for their citizens. This was the case in Japan, where there are 47 prefectures and 1742 municipalities. The size of Japan’s municipalities varies considerably, with Osaka and Yokohama having a few million and small villages with less than 1000 residents. The varied responsibilities exemplify the classic trade-off between efficiency and effectiveness that are reflected in all organizational questions related to the degree of centralization [5]. Many organizations have found that the creation of an enterprise IS architecture involves working out a design that strives to balance global efficiency with local effectiveness [6]. The extent of the damage far exceeded the capabilities of municipal governments, and it took national and international mobilization of

179


Drive

Definition

Universality

The drive to overcome the friction of information systems’ incompatibilities

Ubiquity

The drive to access information unconstrained by time and space

Uniqueness

The drive to know precisely the characteristics and location of a person or entity

Unison

The drive for information consistency

Table 3. The information drives [8].

Message

Code

I need water

1

I need food

2

I am injured and cannot move

3

:…

…

Table 4. Universal message coding. resources (the U.S. military played a major role). At the same time, municipal governments played a critical role, as they are the agencies closest to the residents. They have firsthand knowledge of the people and resources in their area, they are already at the disaster scene because they live there, and they are familiar with local customs and dialects. Such knowledge is critical for outside rescuers to function effectively. The role of the national government, in our assessment, includes supporting the municipal governments to play their critical role by establishing a national disaster recovery infrastructure and standards. It needs to ensure that as soon as possible after a disaster, municipalities have the minimal resources they need for a local response. As the focus is on sustaining life immediately after a disaster, we concentrate on what is required to create an efficient national frugal IS that can be deployed to support effective local relief.

FOUNDATIONS OF A FRUGAL NATIONAL EMERGENCY IS The four u-constructs [7, 8] are a basis for establishing system design principles (Table 3). We consider each of the drives, starting with the universality construct because settling on a common communication platform is the foundation for satisfying the other u-constructs. Universality — A frugal IS must be compatible with existing systems that disaster victims are likely to possess and should require them to acquire minimal new technology, if any. Smartphones, which are advanced cellphones and becoming very common, are portable, battery powered, often within the

180

owners’ easy reach, incorporate a GPS, and have the capability to run apps. Because cellphone adoption is 96 percent globally, 128 percent in developed countries, and 89 percent in the developing world, respectively [9], it is appropriate to make the cellphone the standard platform for a frugal emergency relief IS. To further enhance universality, all cellphones should have a pre-installed emergency app that can be remotely updated as required. Such an app should be created and maintained by the appropriate national emergency agency. Language is a major form of friction between ISs. People in a disaster zone might well speak a different language than some relief workers. Thus, we envisage an emergency app that transmits a code (Table 4) as well as the GPS location and the sender’s phone number. A code can readily be converted into the language of the receiver. A code is frugal in terms of bandwidth requirements. Similarly, there can be another set of frugal messages to support communication with victims. Cellphones require power, but most should have sufficient battery power to continue to operate for a day or so. In addition, we advocate that a hand-powered phone battery charger be added to the recommended household emergency kit. Municipalities should also establish an emergency phone charging resource because their relief workers are likely to use cellphones during the first stage of recovery. Ubiquity — Victims and relief teams need ubiquitous access to essential information wherever they might be in the disaster zone. Because it is likely that essential infrastructure has been destroyed, there is a need to quickly deploy an alternative wireless communication system that is compatible with the victims’ cellphones and can cover the relief area. A cellphone network must provide both coverage and bandwidth to service customers. The GSM specification supports communications up to 35 km, and a frugal IS needs to provide only minimal coverage because frugal messaging minimizes the need for bandwidth. The following provides a summary of some options for supporting continuous cell site connectivity as it pertains to ubiquity. More indepth information is detailed in the Technical Options section. Aerial deployment of cell sites is probably the quickest way to create coverage. The aerial options for extending communication access might include drones, tethered and untethered balloons in the stratosphere, blimps, and aircraft. An aerial cell site offers options for continuous connectivity to an impacted community to meet the operational needs identified at the end of an earlier section. Another alternative to aerial options is to deploy offshore ships or river barges. Such resources can be pre-commissioned and ready to position and activate. Consideration also needs to be given to the power required to operate the temporary cell sites and their connection to the operational remnants of the existing cell network. The goal is to create a minimal ubiquitous access system as soon as possible.



It is important to keep in mind that the disaster area for March 11 covered over 500 km. Sustaining human life is expensive for massive scale disasters.

APPLYING THE PROPOSED FRUGAL IS TO THE GREAT EAST JAPAN DISASTER

Uniqueness — The identity, location, and condition of victims must be established as soon as possible. In addition, there is a need to identify locations that are hazardous (e.g., a radiation leak) or require a rescue team (e.g., a demolished building). Because the cellphone is typically a personal device, there is a one-to-one mapping of a cellphone number to a person. The full identifier, country code plus phone number, needs to be used to allow for visitors. What is often missing, as the disaster demonstrated, is a national or even local government-based, integrated database that records this mapping along with a person’s home and work addresses and some minimal personal (e.g., birth date), family (e.g., details of children), and critical health (e.g., diabetic) data. Such information means that the general and unique needs of a disaster area can be quickly determined (e.g., the number of doses of insulin). While a disaster is often seen as a national tragedy, it is a combination of many individual calamities that frequently have unique needs.

The Internet connectivity compatibility of contemporary cellphones and their widespread adoption makes them the most likely candidate technology for fulfilling the four u-constructs. While hindsight is not the best method of testing a design, it is the best alternative at this point. Consequently, we iteratively refined the design by examining the performance of five key tasks (Fig. 1), and we now discuss how the proposed frugal IS can handle these critical issues. Our proposal focuses on recovery of cell sites, based on our observation that they recovered much faster than other components of the telecommunications infrastructure, such as the WAN that connected municipal government offices.

Unison — Information about victims and where they live can be scattered across multiple databases (automated and manual within the community, in regions, and in national registries), and during an emergency there is often a need to integrate these data. Such integration, however, should be prepared prior to the need, as identified in the prior discussion on uniqueness. The tension between the general need for privacy and the specific needs of a disaster should be accommodated in line with existing or new national laws. Unison also comes into play in another way. Keeping details of victims is critical to managing the recovery. Each relief station needs to keep track of who is housed where, who is receiving assistance, and so forth. Relying on paper records can result in a lack of information consistency. Written names can be miskeyed, and databases then become unreliable. The recording process needs to be fast, efficient, and designed to maintain unison across recording events as victims move. The cellphone is the key to maintaining unison because it is a person’s electronic identifier. The emergency app needs a feature that transmits (e.g., Bluetooth or NFC) its phone number to a rescue worker’s cellphone in the field or at a relief center. Each relief worker’s list of phone numbers and coordinates could be uploaded automatically to a computer system. Creating a single integrated database and ensuring consistent identification of entered data is fundamental to maintaining unison. In the case of an emergency, the key identifier is the phone number of the victim. The lessons from March 11 underscore the importance of unison, and this is perhaps the uconstruct that should be pursued with the most vigor. However, without the other constructs in place, it is not feasible to have unison.


Our proposal focuses on recovery of cell sites, based on our observation that they recovered much faster than other components of the telecommunications infrastructure, such as the WAN that connected municipal government offices.

CONFIRMING THE WHEREABOUTS AND SAFETY OF RESIDENTS From the municipal governments’ perspective, confirming the whereabouts and safety of residents requires recognition of the surviving residents and communication of information to the outside world As for the recognition of the surviving residents, if the cell system is operational, residents can send identifying messages from their cellphones. If the system is not operational, identification can be achieved by registration of residents’ IDs at evacuation centers or other facilities set up by the municipal government. This can be accomplished by setting up a local communication link between cellphones. We strongly suggest that municipal governments collect residents’ cellphone numbers as well as GPS generated location information in addition to conventional name, address, gender, and birthdate. This will help to maintain the uniqueness of identification and database unison. Thus, preparation for integration of official residential data and carrier user data will be essential, but it will likely require authorizing legislation in many countries. The collaboration of citizens with the public sector is thus an important element in applying frugal systems thinking in disaster readiness planning. We also need to consider the case of people who do not have a cellphone with them or do not own one (e.g., children). Those in the first category should be able to remember their phone number, and this can be manually captured. The second category can be identified by adding a digit to the phone number (e.g., +1 999-999-9991 for the youngest child). To prepare for those who have neither a phone number, a cellphone equipped relative, or guardian (e.g., a solitary elderly), we could allocate a set of unused numbers to evacuation centers in advance for identifying such people. As for communications to the outside world, the ideal situation is to have cell sites back in operation quickly. If this is not accomplished, we need to consider alternatives to rapidly create a low-bandwidth ubiquitous communications net-

181


Cloud computing Citizen’s data 4) Create evacuee lists City office

Evacuation center 5) Issue disaster victim certificates (e.g., by QR 2) Open evacuation centers,codes) identify individuals, and record arrivals and departures (e.g., by Bluetooth and NFC) 1) Confirm whereabouts and safety (using phone numbers and GPS)

3) Transport relief goods and update road conditions (M2M)

Figure 1. Frugal design in disaster relief operations.

work. Alternative means of connectivity will be discussed later.

ESTABLISHING AND OPERATING EVACUATION CENTERS Once a municipality has physically established evacuation centers, it needs a frugal IS to record data about who arrives at or leaves a center. Again, the universal nature of the cellphone means it becomes a key device for exchanging essential data between survivors’ and relief workers’ cellphones. An advantage of this approach is that it allows offline local processing when the network to the outside world is not operational. Stored data can be transmitted to the integrated database as cell sites become operational. Local centers should be able to draw on this database using a preset small number of queries that require minimal ICT resources. As power and communication systems are established, there must be a smooth migration to computer-based, rather than phone-based, information systems.

TRANSPORT AND MANAGEMENT OF RELIEF GOODS In the Great East Japan Earthquake recovery process, a lack of information caused a massive mismatch between the supply and demand of relief goods. Not only were there shortages, but there were also instances where relief goods were delivered but rotted because the demand was already met. Perhaps the accuracy requirement is lower than for survivor identification, but it is critical to have unison and uniqueness in the identification of the need for and supply of relief goods. A cellphone-based system would help to more closely match the local needs with supply capabilities. Requests accompanied by the ID of phones that dispatched the request and location data would be useful in tracking how demand is being filled. If the reception of goods could be registered by the same cellphone, redundant supplies should be reduced and diverted to where they are needed. The same applies to transportation resources. For example, as disruption of the road system is

182

likely, those delivering supplies need a separate frugal IS to be able to report failures in the road network and get advice on new routes. The same u-constructs can guide design of this system. We imagine a system where a copy of the existing road network is dynamically updated with the latest availability data and used to create dynamically new routes. Cellphone-based systems can be used to report usability of road systems from the ground up. Use of such citizen-generated information (CGI) will be of critical importance in the design of future communities.

SUPPORT OF EVACUEES AND CREATING EVACUEE LISTS The power of a GPS enabled cellphone-based system is that the data on a citizen’s whereabouts and condition can be collected in the field as well as at evacuation centers with standalone processing. Thus, rescue teams might well include a data officer, carrying a high-capacity battery, who can identify each evacuee’s needs before they reach a relief center so that the center has time to gain some advance warning of the types of support required. In an emergency, an extra few minutes can be lifesaving. The actions taken to confirm the whereabouts of citizens become the foundation of evacuee lists. As these lists would be transmitted to the cloud as soon as bandwidth was available, they could be made generally available, with appropriate privacy safeguards, to inform relatives. Compilation of the victim database is a sensitive operation. One should expect, and try to meet, the many inquiries from the outside world desperately wanting information about relatives and friends. At the same time we should be aware of the privacy concerns with such a database. The Japanese privacy law exempts databases for disaster relief from its regulations. At the same time, ambiguity surrounding the law forced each municipal government to exercise its judgment and prevented them from taking unified action. As a result, one city publicized an evacuee (survivor) list on its home page, while a neighboring city would not even answer telephone inquiries. There is a critical need for unified guidelines to prevent such fragmented actions. Cloud computing could be an effective infrastructure to house a critical database [10]. It would be particularly useful in integrating the numerous databases that might be compiled in evacuation centers. Municipal governments are already taking action to adopt a public cloud to share costs. However, as cloud computing assumes connectivity, its reliability would likely be compromised because of infrastructure damage immediately following a large-scale disaster. Thus, a more local solution is required for situations immediately after a disaster. Data created by local solutions should be transferred to cloud systems when connections are restored.

ISSUING OF DISASTER VICTIM CERTIFICATES The proposed frugal IS captures essential data about each victim and tracks movement between evacuee centers. In effect, it creates an audit trail that can be used to distinguish a genuine



victim from a fake one. Once a person’s right to a certificate has been verified, this could be issued electronically (e.g., send a QR code to their cellphone) as well as recorded in the database for subsequent reference. The issuing of victim certificates illustrates the power of frugal thinking: use what the victim is very likely to already have and avoid additional resources (e.g., printers). Of course, there needs to be a workaround for those who do not have a phone, and in this case a possible frugal solution could be based on small printers that can generate a paper-based QR code.

*Individual

Of course, there needs to be a

WiFi, direct satellite access

workaround for *Key location (evacuation center) : hubs

those who do not have a phone, and

Satellite phones / Internet Aerial (airplane / balloons) solutions wire connection

in this case a possible frugal solution could be

*Outside

based on small printers that can

Figure 2. Emergency network connectivity.

TECHNICAL OPTIONS The proposals presented above assume nondisruption or rapid recovery of the cellphone network. While we observed the relative robustness of the cellphone network compared with terrestrial networks, we should nevertheless assume that the network will fail and consider alternatives. Assuming local cell sites are not operational, we need to think about: • Connectivity between local hubs (e.g., evacuation centers and city offices) and outside • Connectivity between the hub and the individuals’ devices (e.g., cellphones) (Fig. 2) First, likely options for connectivity between hubs and outside are: • Satellite phone/Internet services • Aerial (airplane/balloons) solutions • Wired connection Satellite Internet services (we distinguish IPonly satellite Internet services from conventional voice-centric satellite phone services) would be the most realistic candidate, at least for the time being, albeit with limited capacity. In the case of the Great East Japan Earthquake, ICT support operations organized by the Japan Electronic and Information Technology Association (JEITA) started to restore information systems one month after the disaster [3]. In the three months of its operation, 95 requests/ operations for relief were conducted. While the primary effort was directed at delivery of equipment, connectivity support, such as satellite Internet access, was provided in the most severely hit areas along the coast. In those areas, the trunk cable among cell sites was hit so hard that it was not realistic to restore the system. Instead, portable satellite Internet Earth stations were brought in to enable connectivity to isolated locations, and individual devices were connected using WiFi. Satellite Internet Earth stations can be purchased for less than US$3000/unit with running costs of US$50/mo (1 Mb/s, 512 kb/s) [11]. This is, we assess, an affordable price for many facilities to install for daily use. It is a realistic way to restore connectivity, and likely to be more cost effective and faster than trying to restore trunk lines among mobile phone cell sites [12]. Another possibility is the Iridium (satellite phone) commercial service for which handheld devices are a little more than US$2000. While this is comparable to satellite Internet services, there is a major difference in that the usage charge is more than US$1.5/min (in addition to


generate a papera base charge of about US$50). This means local communities are unlikely to regularly use this technology, which we consider to be a critical factor in evaluating a technology’s usefulness immediately following a disaster. In a crisis situation, unfamiliar technology is an impediment to action. Specifications for spatial connectivity of the aerial cell site will depend on the type of communication device deployed, such as an aircraft or balloons launched for ground coverage. For example, the E-3 Sentry U.S. Airborne Warning and Control System (AWACS) can provide up to 320 km of coverage to facilitate ground to air communications.1 Alternatively, Google’s “Project Loon” proposes floating balloons 18–27 km above the Earth’s surface to cover an area 40 km in diameter to enable connection of hundreds of cells at speeds comparable to third generation (3G). 2 Such balloons offer a faster connection than currently available via satellite Internet. A box of electronics is carried by the balloon, and contains GPS, sensors, Internet antennas, and 100 W of solar power, which runs the electronics and keeps an onboard battery charged for four hours of use at night. 3 The balloons are designed to stay aloft for 100+ days. Additionally, “for balloon-to-balloon and balloon-to-ground communications, the balloons use antennas equipped with specialized radio frequency technology. Project Loon currently uses ISM bands (specifically 2.4 and 5.8 GHz bands) that are available for anyone to use.”4 The system can be equipped with industrial, scientific, and medical (ISM)/ GSM cellular RF antennas, which can provide WiFi capability. As this technology is in the experimental phase, it is currently difficult to determine costs. It would have required about 12 balloons to cover the long rupture zone areas of the Great East Japan Earthquake. Similar prototypes have been tested for ad hoc communication systems in lower and more controlled implementations. For example, SKYMESH, an urgent communications network backbone, supports balloons 50–100 m above the ground. It takes advantage of good line of sight, low interference, and long transmission range for collecting disaster area information for rescue recovery and survey purposes [13]. It should be noted that with aerial solutions, complex legal issues regarding airspace access and control can arise. Nevertheless, there is a

based QR code.

1

Last accessed 17 Aug., 2013, at http://www.fas.org/man/d od-101/sys/ac/e-3.htm 2 “Google Project Loon, Balloon-Powered Internet for Everyone”, last accessed 17 Aug., 2013, at http://www.google.com/lo on/ 3

Last accessed 17 Aug., 2013, at http://www.greenbiz.com/news/2013/06/22/ google-harnesses-windsolar-spread-interneteverywhere-balloons?mkt_ tok=3RkMMJWWfF9ws RokuK%2FNZKXonjHpfsX56%2BwpXaCylMI%2F0ER3fOvrPUfGjI 4DTctqI%2BSLDwEYGJlv6SgFSLHEMa5qw7gMXRQ%3D 4

Last accessed 17 Aug., 2013, at http://readwrite.com/2013/ 06/19/a-handy-guide-togoogle-project-loon

183


An infrastructure based on frugal IS principles should allow officials and citizens to respond creatively in disaster situations. We believe such possibilities for autonomous and inventive solutions will make systems highly resilient to disasters.

strong argument for deploying balloons in dire situations to meet immediate needs to sustain life. Regardless of the choice of connectivity, communication traffic control becomes a major problem in a disaster situation. Ordinary network structures may be disrupted with damaged cables and cell sites. Without adequate control, weakened networks tend to be overwhelmed by higher than usual post disaster traffic. A geographically distributed operation support system (OSS) makes it possible to manage the network to optimize the routing operation from as far as 600 km [14]. Its design goal is to resume operations within 10 minutes after a disaster. This system could be a resilient network at lower cost than a standard dualization-based backup system. Second, likely candidates for connectivity between hubs and individuals are WiFi and direct access to the satellite. Another proposal is a shelter communication system (SCS), which consists of a server and PCs for connecting to the Internet [15]. An SCS could fill information gaps within an evacuee center. It could also supplement municipal government operations because it is designed to conserve bandwidth. Power supply, however, will likely be a major problem for an SCS, as in the case of the Great Japan East Earthquake. Cellphones are less of a problem with regard to charging. The overarching criterion for the usefulness of a potential technology for immediate disaster relief is whether it has the functionality and cost structure to make it part of a person’s daily life. Municipal governments are under severe budget constraints, and they cannot assume large costs for technologies that are used only on very rare occasions. Our analysis shows that technologies with pricing structures that deter daily use have limited success in disaster situations. This implies there is a need to use a wireless network and associated devices not only as an emergency backup but also in normal daily operations. While recognition of the importance of wireless has become stronger among the municipal governments after the March 11 disaster, its adoption by municipal government systems has been slow. We need technical, legal, and organizational measures to allow smoother introduction of wireless systems in municipal government information systems.

CONCLUSION A resilient information system requires minimal power resources, familiarity with communication technology and tools, and versatile applications. Assuming the issues raised previously can be overcome, we propose the following design guidelines to build a resilient municipal government ICT system based on IS frugality concepts. • Prepare disaster resilient and ubiquitously available networks as soon as possible. • Design universal infrastructures and applications. • Instead of government prepared terminals, make use of citizens’ own devices that are numerous (unfamiliar devices especially prepared for disasters were not used).

184

Specifically, use mobile devices to identify individuals and secure uniqueness. • Promote public and private sector collaboration for data unison. It is worth noting that no single solution solved the problem for the disaster of March 11, but we believe a cellphone-based solution, while maybe not the complete answer, will be a major advance in disaster relief. There is, however, a need for compatibilities at various levels so that various frugal systems (including handwriting) can be integrated to serve the ultimate task of effective and efficient relief. Governments should invest in the resources to quickly create the infrastructure to sustain a frugal IS. The success of the system is dependent on an individual’s comfort and familiarity with the disaster reporting app and understanding that their cellphones’ capabilities will be constrained to basic messaging. People will need to be educated and given the opportunity to practice with the frugal tools in the “peace time” before a catastrophic event. This testing will allow citizens to understand their responsibilities and provide load tests of the system’s infrastructure. A disaster creates massive disruption of human life, physical structures, and ICT assets. Information about the state of this perturbance is usually critical if the inevitably scarce rescue and recovery resources are to be used effectively. A frugal IS is an initial step toward providing such information and should be a central part of both national and local disaster plans. The authors caution readers that this article uses the Great East Japan Earthquake as the basis for evaluating the adequacy of our proposal, and it has not been benchmarked against every conceivable disaster. At the same time, we are reasonably confident that the core elements of a frugal IS design would be applicable in most other disasters. An infrastructure based on frugal IS principles should allow officials and citizens to respond creatively in disaster situations. We believe such possibilities for autonomous and inventive solutions will make systems highly resilient to disasters. Japan’s experience in building and deploying such a frugal IS could greatly enhance disaster recovery across the globe.

REFERENCES [1] R. T. Watson, K. N. Kunene, and M. S. Islam, “Frugal IS,” Information Technology for Development, vol. 19, issue 2, 2013, pp. 176–87. [2] Y. Shibata et al., “Problem Analysis and Solutions of Information Network Systems on East Japan Great Earthquake,” 26th Int’l. Conf. Advanced Information Networking and Applications Wksps. Proc., 2012, pp. 1054–59. [3] Secretariat of the Great East Japan Earthquake Information & Communication Technology Support Team, “Great East Japan Earthquake Information & Communication Technology Support Team Activity Report,” Tokyo, Japan, 2011. [4] M. Sakurai and J. Kokuryo, “Municipal Government ICT in 3.11 Crisis: Lessons from the Great East Japan Earthquake and Tsunami Crisis,” Berkman Center for Internet & Society in Partnership with Keio University, Cambridge, MA., 2012. [5] C. A. Bartlett and S. Ghoshal, “Managing Across Borders: New Strategic Requirements,” Sloan Management Review, vol. 28 no. 4, 1987, pp. 7–17. [6] H. A. Smith, R. T. Watson, and P. Sullivan, “Delivering Effective Enterprise Architecture at Chubb Insurance: A Case Study,” MISQ Executive, vol. 11 no. 2, 2012.



[7] R. T. Watson et al., “U-Commerce: Expanding the Universe of Marketing,” J. Academy of Marketing Science, vol. 30, no. 4, 2002, pp. 333–47. [8] I. A. Junglas and R. T. Watson, “The U-Constructs: Four Information Drives,” Commun. AIS, vol. 17, 2006, pp. 569–92. [9] ITU, “The World in 2013: ICT Facts and Figures,” Geneva, Switzerland, 2013. [10] M. Pokharel et al., “Disaster Recovery for System Architecture using Cloud Computing,” 10th Annual Int’l. Symp. Applications and the Internet Proc., 2010, pp. 304–07. [11] Ministry of Internal Affairs and Communications, “Report of the Usage of Satellite Internet in a Disaster Situation,” Tokyo, Japan, 2011 [12] K. Kataoka et al., “LifeLine Station: A Quickly Deployable Package for Post Disaster Communications,” Internet Conf. Proc., 2009, pp. 41–47. [13] H. Suzuki et al., “An Ad Hoc Network in the Sky, SKYMESH, for Large-Scale Disaster Recovery,” 64th VTC Proc., 2006, pp. 1–5. [14] Y. Takeuchi et al., “The Proposal of Geographically Distributed OSS Against a Great Earthquake,” 13th AsiaPacific Network Operations and Management Symp. Proc., 2011, pp. 1–8. [15] K. Mase, “How to Deliver Your Message from/to a Disaster Area,” IEEE Commun. Mag., vol. 49, no. 1, 2011, pp. 52–57.

BIOGRAPHIES MIHOKO SAKURAI ([email protected]) is a Ph.D. student at Keio University’s Graduate School of Media and Governance where she studies effective ways of using ICT in Japan’s municipal governments. She received her Bachelor’s degree in policy management from Keio University in


2005. She has worked for the Japan Newspaper Publishers & Editors Association. She acquired a Master’s degree from the Graduate School of Media and Governance of Keio University in 2011. R ICHARD W ATSON ([email protected]) is the J. Rex Fuqua Distinguished Chair for Internet Strategy in the Terry College of Business at the University of Georgia. He is a former President of the Association for Information Systems and the current Research Director for the Advanced Practices Council of the Society of Information Management. In 2011, he received the Association for Information Systems’ LEO award, which is given for exceptional lifetime achievement in information systems. CHON ABRAHAM ([email protected]) is the Wakefield Distinguished Associate Professor of Business at the College of William and Mary. She received a B.S. and a commission from the United States Military Academy at West Point, an M.B.A. from Old Dominion University, and a Ph.D. in management information systems from the University of Georgia. She worked as a systems analyst for American Management Systems, and is a cyber and information systems officer in the U.S. Air Force Reserve. JIRO KOKURYO ([email protected]) is a vice president of Keio University as well as a professor in its Faculty of Policy Management. His research and teaching interests center on the development of business and social models that maximize the benefits of information technologies. He is a former President of the Japan Society for Management Information. He received his D.B.A. degree from Harvard University in 1992. Prior to joining Keio University, he worked for Nippon Telegraph and Telephone. His public service includes committee membership in the Prime Minister’s IT Strategic Headquarters.

185