Communications of the ACM
Viewpoint
Responding to the Unexpected
This month marks the second anniversary of September 11, 2001. The events that day represent a fundamental shift in the type of disaster we recognize as even being possible. However, these years have not seen a corresponding change in the resources available to government and first responders having to deal with the immediate aftermath of such attacks. They should have.
The events of 9/11 were a radical departure from the past in that they were unexpected by the people charged with dealing with them. It is not simply that their precise nature and timing were a surprise; they were entirely outside the range of the planned capabilities of the organizations mandated to deal with them. Developing the ability to respond to the unexpected continues to require a major interdisciplinary research effort.
The response issue concerning unexpected events is not simply a lack of preparedness. Government agencies have in place significant mechanisms for managing a range of natural and man-made disasters, from earthquakes, to hurricanes, to disease epidemics, to the accidental release of toxic or radioactive substances into the environment. Considerable investment has gone into research to address such catastrophes. But assessment and recovery efforts are extremely problematic in the aftermath of events so rare and damaging they outstrip preparations, plans, and human and material resources. A society of limited resources cannot afford to prepare for each and every conceivable disaster. Simply trying to identify each scenario would be prohibitively expensive, as well as being doomed to failure. Unplanned-for, low-probability, high-impact events with significant systemwide effects will continue to occur, despite our best efforts to prevent them.
While it may seem paradoxical to attempt to prepare for the unexpected, the current state and promising near-term future of IT make it possible to create policies, build a general-purpose infrastructure, and develop general capabilities that could be adapted instantly to respond to any threat. Society can create a foundation on which an effective response is quickly constructed. But doing so requires a major enterprise; the National Science Foundation recently estimated that $3 billion over 10 years is required to fund interdisciplinary efforts, from individual projects at the low end, to local, regional, and national centers capable of developing and supporting large-scale systems and testbeds. However, in the interests of our common future it is imperative that the funding agencies develop the means to support such an enterprise and that significant segments of the research and development communities commit to the large-scale, sustained effort required to make it happen.
Research Agenda
Research is thus required in a number of disciplines; here, we focus primarily on critical advances in IT.
Encyclopedic digital collections. A first priority is to create a digital collection of information on geography, environments, resources, buildings, computational facilities, and potential response personnel and organizations, together with software systems that can locate answers to pertinent questions. Research is required to develop general tools that could analyze very large stores of data, identify patterns, reason, extract and summarize content in natural language and in graphical and other forms, translate text between languages, incorporate incoming data into the existing store, and answer questions. Given any task, relevant information and instruction in how to use it must be extracted automatically. The knowledge need not (and surely will not) be centralized in a single location, and access to distributed information and data management facilities must be transparent to users. Privacy issues must also be resolved before this kind of store of knowledge can be created.
A Grid of unlimited computation. A nationwide network of computation, data, and services supporting utilization of any resource needed during a response must be assembled. Grid technology is being developed to support the sharing of computation and resources distributed over the Internet. Also needed are secure and transparent means of accessing and combining these resources.
Rapidly deployable sensors and effectors. The related technology ranges from microsensors to Earth-observing satellites, and from simple actuators to autonomous robots. Instantly deployable, self-configuring versions must be created. Available today are devices capable of detecting motion, heat, light/images, sound, pressure, the presence of metal, and much more. Research focuses on developing micro-scale "machines" and reconfigurable robots. We must also be able to quickly deploy the sensors and effectors so they network autonomously among themselves and communicate with controllers outside the crisis zone. These units would gather data and, functioning autonomously, convey firsthand information to emergency managers who could issue additional commands remotely, facilitate search-and-rescue missions, and work in teams with and support human responders. However, as with the other technologies, relevant privacy concerns must be respected.
Society cannot afford to prepare for every eventuality, but it can create a foundation on which an effective response is quickly constructed.
A pervasive, secure communications infrastructure. Data, computers, and networks used in the response process must all be rendered secure from eavesdropping and sabotage and made less susceptible to disruptions and loss of connectivity. A concerted effort to develop new methods to protect computer systems from intrusion and sabotage and enable recovery from such disruptionspossibly utilizing extended virtual organizationswould ensure that emergency response teams operate in safety, free from surveillance and malicious interference. New security initiatives must also cover wired and wireless networking and communication for both speech and data. New technology is needed to allow networks to be readily deployed, enabling networks to self-configure and self-repair.
Integrated analysis, fusion, and learning. Computer-aided learning and training must be embedded within systems and made available to users as needed. Urgent tasks and a wealth of incoming data make it impractical for individuals pursuing their duties to spend too much time trying to discover the information they need or that otherwise affects their ability to perform. Learning and training technology must be adapted to determine which people need to perform their jobs at which time and ensure that information is tracked and delivered when appropriatejust as instructional systems identify gaps in knowledge relevant to individual students and act to make it available. In addition, training to use these systems must be embedded in the systems themselves and made available as needed.
These technologies can be combined to provide visionary capabilities that could become central to our common ability to respond to the unexpected. For example, we would be able to create comprehensive regional models, real-time sensor-updated models of significant geographic regions, even entire cities, including subsurface properties, utilities, transportation, structures, population, weather, and more. We would thus have the ability to run what-if simulations to evaluate proposed response procedures, discover potential weaknesses, provide integrated situation awareness to response personnel, and use the technologies as testbeds for training. These models would also have tremendous potential in urban planning, public policy setting, social science research, and education.
Virtual organizations. Uniting geographically dispersed people, software, and hardware systems into flexible, resilient, dynamic, and coordinated teams, they would also be essential parts of the response in emergency situations. Many modern computer systems are far more than tools, reflecting group goals, as well as their own roles in achieving them, and thus can function as active participants. Agents and robots engaging in information discovery, task management, and coordination could participate in and help sustain virtual organizations.
Legal framework. Several of these developments have the potential to unduly infringe on personal privacy and civil liberties. It is imperative that we find ways to protect individual privacy and civil liberties while resolving the jurisdictional and legal barriers that hamper or even prevent necessary and proper information gathering and sharing by governmental and other organizations during unexpected events.
Moreover, laws and technologies should enable selective strengthening and weakening of security and privacy policies during such events by, say, allowing responders to access critical information normally withheld from them or by tightening security on critical facilities when there is a strong probability of attack.
Looking Ahead
Last year, our interest in these matters led to the creation of the Center for Research on Unexpected Events (CRUE), whose mission is to help transform the U.S.'s ability to respond to unexpected events by building on revolutionary IT developments (see crue.isi.edu). Uniting the University of Southern California's existing Digital Government Research Center, Center for Computer Systems Security, Center for Grid Technologies, and Center for Advanced Research in Teaching for Education, CRUE's major research groups cover agents, sensor networks and fusion, situation awareness, geographic information systems, and information integration. External partners from universities, industry, government, and response organizations are developing and deploying the new technologies, including comprehensive regional models, virtual cities, and virtual organizations, to enable response to future unexpected events.
We had previously suggested to NSF the need to develop a research agenda along these lines, and it asked us to organize a workshop to study the matter. Held in February/March 2002 in New York, it was called "Responding to the Unexpected" and attracted participants from universities, government agencies, and business, including researchers from the social sciences, engineering, and computer science. Responders and practitioners represented the White House Office of Homeland Security, Federal Emergency Management Agency, Environmental Protection Agency, and Coast Guard. People involved in responding to the events of 9/11, as well as to earthquakes and system failures (such as nuclear-plant incidents and electrical blackouts), shared their experience and insights. The result was a series of recommendations for research in several key areas, along with possible modalities for developing technology and applications useful over the next five years.
Conclusion
It's worth noting that while the research initiatives contemplated here might be crisis-driven, they could be brought to bear on a range of problems endemic to government and society in general. No one foresaw the explosion of media, art, e-commerce, and education that resulted from the communication infrastructure of the Internet. The new research, technology, and infrastructure might initially be aimed at improving our ability to respond to unexpected disasters. But, fully realized, it could revolutionize the way we address many other societal problems and organize social activity.
©2003 ACM 0002-0782/03/0900 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2003 ACM, Inc.
No entries found