Computing Research in the FY 2000 Budget Request

The following appears as a chapter in AAAS Report XXIV: Research and Development FY 2000.

Highlights

Introduction and Background

The mix of federally supported computing activities is dominated by several high-profile programs and coordinated multiagency efforts, including HPCC and NGI, in which computing, computational, information, and networking research are tightly interwoven. They are the present manifestation of a complex partnership among government, universities, and industry that has evolved over the past 50 years to advance the frontiers of information technology.

Within the context of this partnership, strategic federal investments in computing research drive precompetitive innovation and stimulate revolutionary advances in information technologies; improve the accessibility, robustness, reliability, and security of information infrastructure; and provide high-end computational and networking resources for government and university scientists working on problems of national importance.

It is often noted that federal research investments were responsible for the creation of the Internet. The Internet is just one facet of an information technology industry that accounts for more than $500 billion of the annual U.S. economy. Businesses that produce computers, semiconductors, software, and communications equipment have been responsible for one-third of the overall growth in U.S. production since 1992, creating millions of new high-paying jobs. Moreover, as prices in these sectors have declined more rapidly than most, information technologies industries have been keeping inflation a full percentage point lower than it otherwise would be.

Current Policy Environment

In the legislation that created HPCC and NGI, a private-sector panel was authorized and charged with assessing the overall federal investment in information technology (IT) R&D. The President's Information Technology Advisory Committee (PITAC), which issued its final report in February 1999, determined that federal support for IT R&D is inadequate and too focused on near-term problems; long-term fundamental IT research is not sufficiently supported relative to the importance of IT to the United States' economic, health, and other aspirations; critical problems in computing are going unsolved and the rate of introduction of new ideas is dangerously low.

PITAC recommended that the federal investment in information technologies R&D be increased by $1.37 billion over the next five years and that a strategic initiative be created to support fundamental research that will lead to breakthroughs and new capabilities to serve the growing demands on information technologies, given that IT is rapidly being incorporated into many sectors of American society.

In response to these recommendations, the Administration is developing plans for an Information Technology for the Twenty-First Century initiative that would be funded at $366 million in FY 2000. It calls for an adjustment in how policymakers and R&D officials regard computing research, which is often thought to be fully covered by the HPCC program and its successors. HPCC activities emphasize computational research-providing infrastructure and applications to solve computational problems in various fields of science and engineering. The IT initiative would complement HPCC by addressing critical questions in computing research, the answers to which will have impact on a broader range of information technologies, not just high performance or high-end computing.

The PITAC study identified many of these critical issues: making software more reliable and less costly to develop and test; designing infrastructure that is easily scaled up to serve more users and uses as they emerge; and developing innovative architectures and software strategies to utilize the capacity of high-end computers more efficiently.

When it comes to computing policy, the Congress often focuses on high-end hardware issues-where machines are located, from what sources they are purchased, and who has access to them. A current concern is that the computational resources available for civilian research are not on par with those available to researchers working on defense-related problems, in particular, those associated with the Department of Energy's Accelerated Strategic Computing Initiative (ASCI). The ASCI program, part of DoE's nuclear stockpile stewardship efforts, involves the development and deployment of terascale (trillions of calculations per second) computing resources necessary to simulate nuclear explosions to reduce the need for live testing.

To redress the disparity, the IT initiative includes a high-end computing and applications component that would support the acquisition of terascale computational resources for civilian research. Congress will no doubt take an interest in the distribution of these resources among the participating agencies as well as among the laboratories or centers in which the machines would be located.

Budget Request and Priorities

Information Technology Initiative (IT2 or IT-squared): The new initiative would include three components: long-term information technology research that will lead to fundamental breakthroughs in computing and communications; advanced computing for science, engineering, and the nation; and research on the economic and social implications of information technologies, including workforce issues. Six federal agencies have requested funds to participate in the initiative.

FY 2000 IT2 Agency Budget Proposals
Fundamental Information Technology Research Advanced Computing for Science & Engineering Social
Implications &
Workforce
Programs
Total
NSF 100 36 10 146
DOD 100 -- -- 100
DOE 6 62 2 70
NASA 18 19 1 38
NIH 2 2 2 6
NOAA 2 4 -- 6
TOTAL, IT2 228 123 15 366

(amounts in millions, current dollars)
A new senior management team has been formed within the National Science and Technology Council to set policy and coordinate initiative activities. The team is being supported by a working group charged with preparing research plans and budgets for the entire effort. High priority has been placed on ensuring that there is a sound and balanced research portfolio; that agency funds are distributed in an open, competitive process aimed at supporting the best possible ideas; and that resources are well-coordinated and leveraged for the mutual benefit of the participating agencies and their constituencies.

High Performance Computing and Communications (HPCC) and Next Generation Internet (NGI): The FY 2000 budget document includes a list of spending plans for selected computing, computational, and communications research programs in various agencies; the aggregate account is identified as HPCC and includes NGI funding.

HPCC Agency Budgets
FY 1998
actual
FY 1999
estimate
FY 2000
proposed
Commerce 20 27 27
DOD 220 168 207
DOE (Civilian) 115 126 116
DOE (Defense) 374 484 543
Health & Human Services 98 111 115
EPA 3 4 4
NASA 120 93 136
NSF 265 301 314
TOTAL, HPCC 1,215 1,314 1,462

(amounts in millions, current dollars)

However, for purposes of interagency coordination the HPCC program has been succeeded and subsumed by a broader activity referred to as Computing, Information, and Communications (CIC) R&D and coordinated by the National Coordination Office that was established in the HPCC legislation.

Many though not all of the HPCC-labeled activities are included in the CIC set of activities, for which a cross-cut report is typically issued in the fall. The CIC program component areas include High End Computing and Computation, Large Scale Networking, High Confidence Systems, Human Centered Systems, and Education, Training, and Human Resources. The Large Scale Networking component includes the NGI program.

National Science Foundation (NSF): The NSF is serving as the lead agency in the IT initiative, which means, for instance, that the Director of the Computer and Information Sciences Directorate (CISE) is chairing the working group mentioned above. The NSF's investments under the IT initiative would include a $110 million research component in CISE and a $36 million infrastructure component funded through the Major Research Equipment line-item. The latter would provide machine capabilities in the multi-teraflop range, permitting researchers to address problems of scope and scale that are inaccessible with current computer systems. Development locations for the machines will be awarded on a competitive basis.

The CISE research component would encompass broad, thematic, large-scale, long-term, basic computing research. Topics would include major software challenges; human-computer interactions and information management; broadband tetherless communications to help enable new technologies such as telemedicine and distance learning; understanding, modeling, and predicting the behavior of networks; and computational research. CISE plans to use $80 million for individual and team research projects and $30 million to establish new IT research centers.

The overall CISE budget, which would grow by 41.5 percent to $422.5 million in FY 2000, would also include an increase for existing HPCC/CIC activities, in particular the Partnerships for Advanced Computational Infrastructure.

Department of Defense (DOD): DoD has requested $100 million in FY 2000 funding for participation in the IT initiative, including $70 million for focused research programs at the Defense Advanced Research Projects Agency (DARPA); $20 million for a new Advanced Research and Development Activity (ARDA); and $10 million for fundamental IT research through the DoD-wide University Research Initiative. DoD's strategies for IT R&D follow from broader defense and warfighting strategies, developed at higher levels in DoD and the Joint Chiefs of Staff, which identify the growing importance of information superiority to U.S. objectives.

DARPA's participation in the initiative would be consistent with its overall investment strategy of concentrating on high-payoff ideas and technologies with vision and focus. Research areas would include software; human-computer interaction and information management; scalable networks; and high-end computing. ARDA is a joint effort of the Defense Department and the intelligence community to support long-term research on problems and enabling technologies relevant to intelligence and information security.

Department of Energy (DOE): DoE's FY 2000 budget would include an increase for the ASCI program as well as new funding for a civilian counterpart, the Scientific Simulation Initiative (SSI), which would be DoE's contribution to the IT initiative. The objectives of SSI would be to: design, develop, and deploy computational simulation capabilities to solve scientific and engineering problems of extraordinary complexity; discover, develop, and deploy crosscutting computer science, applied mathematics, and other enabling technologies; and establish a national terascale distributed scientific simulation infrastructure.

DoE would initially focus on two major simulation projects that are critical to the agency's mission, have urgent deadlines, are of high scientific impact, and are well positioned to take advantage of terascale computing capabilities: global systems and combustion systems. A number of other applications in basic science that fulfill these criteria, e.g., genomics, would also be initiated.

DoE's efforts in crosscutting and enabling technologies would be expected to both drive and profit from fundamental research in information technology. The hardware strategy would be driven by the applications requirements and would be based on the acquisition of a balanced system of advanced computers for computational methods and software development as well as the demanding applications described above.

National Aeronautics and Space Administration (NASA): With its IT initiative funding, NASA would support activities in three areas: fundamental intelligent systems research focused on advances in software, human-centered computing and information management, and high-end computing; research in applications and use of terascale infrastructure; and training and education efforts. The activities would be designed to serve a number of NASA objectives and missions. For instance, they would include research on automated reasoning to enable greater use of robotics in space exploration; computing research to improve technologies for understanding and managing large and distributed streams of data, such as those generated in earth observation activities; and development of an Intelligent Synthesis Environment testbed for reducing the costs of space transportation.

National Institutes of Health (NIH): The NIH's participation in the IT initiative underscores the increasing interdependency of biomedical research and computing. Its IT activities-in software and algorithm R&D, high-end computing, and workforce training-would be designed to meet the needs of biomedical research and researchers, especially in bioinformatics, imaging and statistical analysis, simulation methods for implementing realistic models of biological systems, and processing large data sets.

National Oceanic and Atmospheric Administration (NOAA): Computer modeling is central to NOAA's missions in weather forecasting and climate research. With its initiative funding, NOAA would address important challenges in the development and implementation of climate and weather applications for advanced computer architectures, pushing the state of the art in the use of advanced high-speed computing, visualization, and data communications for these applications. By improving their models and making them work with the latest technologies , NOAA would expect improvements in its ability to forecast hurricanes, tornadoes, medium- and long-term climate, as well as in routine weather forecasting.


written by Lisa Thompson, CRA Director of Government Affairs

Document last modified on .