Appendix B: Applications and Information
and Communications R & D

B.1. Introduction

The ultimate application of information systems to meet societal needs, such as health care, the environment, and education, is an important long-term goal of information and communications research and development. Pilot implementations, testbeds, and demonstrations of these applications areas provide a crucial vehicle to test and perfect advanced information, computation, and communications technologies. Collaborative testbeds and demonstrations are an effective vehicle for transferring to the private sector reference architectures, validated interfaces and protocols, and proven technologies.

The following examples, drawn from three representative domains (health care, environment, and education and training) are currently being demonstrated within the HPCC Program, as described by such documents as [OSTP 92, 93b, 94a, 94b]. A description of the Program's successes was entered into the Congressional Record as a supplement to the May 10, 1994, testimony before the Science Subcommittee of the House Science, Space, and Technology Committee.


B.2. Health Care

Many modern health care objectives readily lend themselves to solution through advanced information systems. Two classes of advanced diagnostic imaging systems have revealed good experimental results. The first class involves the display of computed molecular shapes and drug interactions. Scores of hours of supercomputer time are required to model significant proteins and protein receptor sites, and useful drugs have resulted directly from this modeling. The second class involves research diagnostic imaging. This imaging combines geometrical modeling (of an optimal number of radiation treatment sources) that can be done in sufficiently "real-time" to serve the human patient only when high performance, high speed "bursty" communications between cooperating supercomputers can be assured. There is simply no theoretical substitute for testing with actual patterns of computational and user demands.

The use of telemedicine to provide expert medical consultation, diagnosis, and management to rural, remote, frontier, and socially isolated areas is being tested in HPCC-sponsored Biomedical Testbed Networks that depend upon high performance communications. These applications allow us to test realistically such major problems as: prototype computer based patient record systems requirements; models of medical data privacy protection; determining the medical circumstances under which very high resolution clinical medical imaging is and is nor required; and direct comparison between on-line two-way video systems versus much less costly store-and-forward systems. So far, the experiments have yielded for medical research better cure rates and far less morbidity secondary to radiation, and for the industrial partners, successful commercial communication products. These tests make clear that privacy of health data is a critical obstacle. Yet, the major problem is not to devise new technical encryption systems that will ensure absolute privacy. The problem is to create systems within a realistic testbed health care environment that will be acceptable to patients, will balance the need to make records readily available - available faster and in more places than now - and yet will also preserve individual privacy.


B.3. The Environment

As a result of accomplishments in the six strategic focus areas Americans will know more about and better manage their physical environment. On-going HPCC Grand Challenge work regarding Chesapeake Bay pollution and weather modeling are illustrative. The Bay is dying due to nutrient overload - the oyster and crab populations that for hundreds of years were key to the Bay's economy are so depleted that in some years commercial fishing is banned. On a shorter time scale, hurricanes and other severe storms destroy shoreline, businesses, and homes.

With the development and deployment of CIC technologies can come better management of Bay ecology and faster response to natural and other disasters.

Until recently, the understanding of the Chesapeake Bay pollution problem was through a water pollution model that described the deposition of nitrogen and other chemicals in fertilizers. These fertilizers are applied to farms as far away as Pennsylvania and New York State, and produce a runoff that is washed into the Susquehanna and other rivers that flow into the Bay. This model describes the influx of such nutrients and how they are affected by Bay features such as varying depths and salinity. There is also a separate air pollution model that describes air-borne pollution introduced by power plants and automobiles. Until recently, these models were run on some of the most powerful computers but still could only be run separately - no computer had the speed and memory to run them together. That situation changed recently. EPA and other Federal agencies have begun to link these two models using the newest generation of large, scalable, high performance systems.

Early results suggest that a large proportion of the nitrogen in the Chesapeake Bay comes from the Bay's airshed that extends across the Allegheny Mountains. If these results are confirmed, they may alter the decisions made by EPA and the affected states about how to improve Bay water quality. Complementing this Grand Challenge research is R&D on how to present this information to Federal and state decision makers most effectively.

Compared to nitrogen pollution, storms have comparatively short-term effects on the Bay's ecosystem. Residents, workers, and travelers need to know whether they can safely venture forth for a week of commercial or recreational fishing, when they should plant their fields and harvest their crops, or that they should stockpile foods for "the blizzard of the century" such as happened in 1993. Pilots need to know about the storms too, so they can avoid them - and they will be better prepared to fly through them as a result of training on simulators in "virtual environments." When there is a major storm or other emergency, rescuers need to know efficient evacuation routes in order to help the injured through on-site medical assistance and transportation to hospitals.

Modeling the environment and the weather has for decades benefitted and continues to benefit from using the largest and fastest computers available. The globe-spanning information infrastructure is critical to monitoring pollution and weather and to transmitting measurements to advanced data storage and computing facilities and then to researchers, decision makers, and the general public. Training in virtual environments helps search and rescue personnel prepare for emergencies. User-centered systems enable all citizens, including those with special needs, to know about their environment and take appropriate action in case of emergencies. How to build an infrastructure that provides capabilities like these is the subject of CIC R&D.


B.4. Education and Training

The area of Education and Training has numerous constituencies. There are the historical constituencies of K-12, colleges and universities, post-baccalaureate students, and life-long learners. There are also practitioners (medical professionals, environmental decision makers, weather forecasters, farmers, pilots, and so on) who need to keep up with the latest research results and methods. There are researchers who need to become facile in using new tools (such as the Internet, or scalable computing systems, or expensive remote instruments). Last but not least are individual citizens - for example a person who wants to know more about a recently diagnosed medical condition.

The deployment of capabilities developed by CIC R&D will make the education and training of all of these constituencies more easily available and more effective. Students will access information no matter where their schools are located by using the Internet to connect to information repositories such as libraries and museums. Practitioners will use similar tools (they may be faster and connect to more resources); some will use virtual environments for training in handling dangerous, rare, or new situations. Researchers will use even more advanced tools (faster Internet connections to scalable computing facilities, large databases, and expensive remote instruments). For each of these communities, the resources they need will be part of the globe-spanning information infrastructure and will be provided in ways that best suit their special needs.