This document was made by OCR
from a scan of the technical report. It has not been edited or proofread and is
not meant for human consumption, but only for search engines. To see the
scanned original, replace OCR.htm with Abstract.htm or Abstract.html in the URL
that got you here.
Evolving the High
Performance Computing and Communications Initiative to Support the Nation’s
Information Infrastructure
Executive
Summary
Information technology drives many of today's
innovations and offers still greater potential for further
innovation in the next decade. It is also the basis for a domestic industry of
about $500 billion,' an industry that is critical to our nation's
international competitiveness. Our domestic information
technology industry is thriving now, based to a large extent on an extraordinary
50-year track record of public research funded by the federal
government. creating the ideas and people that have let industry
flourish. This record shows that for a dozen major innovations, 10 to 15 years have
passed between research and commercial application (see Figure ES.1). Despite
many efforts, commercialization has seldom been achieved more
quickly.
Publicly funded research in information
technology will continue to create important new technologies and
industries, some of them unimagined today, and the process will continue to
take 10 to 15 years. Without such research there will still be innovation, but
the quantity and range of new ideas for U.S. industry to draw from will be
greatly diminished. Public research, which creates new
opportunities for private industry to use, should not be confused with
industrial policy, which chooses firms or industries to support. Industry,
with its focus mostly on the near term, cannot take the place of government in
supporting the research that will lead to the next decade's advances.
The High Performance Computing and Communications
Initiative (HPCCI) is the main vehicle for public research in information
technology today and the subject of this report. By the early
1980s. several federal agencies had developed independent programs to advance
many of the objectives of what was to become the HPCCI. The program
received added impetus and more formal status when Congress
passed the High Performance Computing Act of 1991 (Public Law 102-194)
authorizing a 5-year program in high-performance computing and communications.
The initiative began with a focus on high-speed parallel computing and
networking and is now evolving to meet the needs of the nation for widespread
use on a large scale as well as for high speed in computation
and communications. To advance the nation's information infrastructure there is
much that needs to be discovered or invented, because a useful -information
highway- is much more than wires to every house.
As a prelude to examining the current status of the HPCCI,
this report first describes the rationale for the initiative
as an engine of U.S. leadership in information technology and outlines the contributions
of ongoing publicly funded research to past and current progress in developing
computing and communications technologies (Chapter 1). It then describes and
evaluates the HPCCI's goals, accomplishments, management, and
planning (Chapter 2). Finally, it makes recommendations aimed at
ensuring continuing U.S. leadership in information technology through wise
evolution and use of the HPCCI as an important lever (Chapter 3). Appendixes A
through F of the report provide additional details on and
documentation for points made in the main text.
A
few examples
CTSS,
Multics, BSD Unix
SDS
940. 360167. VMS
Sketchpad, Utah GM/IBM, LucasFilm
E&S, SGI.
Arpanet, Internet Ethernet, Pup, Datakit
DECnct,
LANs, TCP/IP
Lisp
machine, Stanford Xerox
Alto
Apollo, SUN
Englebart,
Rochester Alto, Smalltalk
Star, Mac, Microsoft
Berkeley,
Stanford IBM 801
Sun. SGI, IBM, HP
Mead/Conway, Mosis
many
Berkeley
Striping.
Datamesh many
Illiac
4, C.tnmp. HPC IBM RP3.
Intel
CM-1, Teradata, T3D
|
Gov't research
Transfer of ideas or people
|
Industry research nm.
|
|
Industry
development El - -
|
$ 1B business
|
|
|
|
|
|
|
|
FIGURE
ES.1 Government-sponsored computing research and development stimulates
creation of innovative ideas and industries. Dates apply to horizontal
bars, but not to arrows showing transfer of ideas and people.
INFORMATION TECHNOLOGY—FUNDAMENTAL
FOR SOCIETY
AND THE ECONOMY NOW AND TOMORROW
Computers, the devices that process information,
affect our lives both directly and indirectly. Today, more than
70 million microcomputers are installed in the United States. and between one-fifth
and one-third of U.S. households have one.' Entertainment, education,
communications, medicine, government, and finance are using computers in more
ways to enhance our lives directly through the provision of
such services as distributed learning and remote banking. Computers are also
used to make essential products and activities cheaper and better: airplanes.
molded plastics, automobiles, medical imaging, and oil exploration
are only a few of many examples. A broader benefit is the $500
billion industry's creation of jobs. taxes, profits, and exports.
Clearly, the uses and applications of information
technology will continue to grow. In fact, the information
revolution has only just begun. Computers will become increasingly valuable to industries
and to citizens as their power is tapped to recognize and simulate speech,
generate realistic images, provide accurate models of the physical
world, build huge automated libraries, control robots, and help with a myriad
of other tasks. To do these things well will require both computing
and communications systems many times more powerful than we have today. Ongoing
advances in knowledge will constitute the foundation for
building the systems and developing the applications that will
continue to advance our quality of life and ensure strong U.S. leadership in information
technology. Strong leadership in information technology in turn supports other
sectors including industry', health, education, and defense by
serving their needs for equipment, software, and know-how.
The Basis for Continuing
Strength—
A Successful Government-Industry Partnership
Federal investment in information technology
research has played a key role in the U.S. capability to
maintain its international lead in information technology. Starting in World
War II publicly funded research has helped to stock the nation's
storehouse of trained people and innovative ideas. But our
lead is fragile. Leadership can shift in a few product generations, and because
a generation in the computing and communications industry is at most 2 years,
our lead could disappear in less than a decade.
Since the early 1960s the U.S. government has
invested broadly in computing research, creating new ideas and trained people.
The result has been the development of important new technologies for
time-sharing, networking, computer graphics, human-machine interfaces. and parallel
computing, as well as major contributions to the design of very large scale
integrated circuits, fast computers and disk systems, and
workstations (see Figure ES.1; see also Chapter 1, Box
1.2 for details). Each of these is now a multibillion-dollar business. From
these successes we can learn some important lessons:
· Research has kept paying off over a long period.
· The payoff from research takes time. As
Figure ES.1 shows, at least 10 years, more often 15, elapse between
initial research on a major new idea and commercial success. This is still
true in spite of today's shorter product cycles.
· Unexpected results are often the most important. Electronic
mail and the "windows" interface are only two
examples; Box 1.2 in Chapter 1 outlines more.
· Research stimulates communication
and interaction. Ideas flow back and forth between research
programs and development efforts and between academia and industry.
· Research trains people, who
start companies or form a pool of trained personnel that existing
companies can draw on to enter new markets quickly.
·
Doing research involves taking risks. Not
all public research programs have succeeded or led to clear
outcomes even after many years. But the record of accomplishments suggests that
government investment in computing and communications research has been very productive.
Government Support of Research Is Crucial
The information technology industry improves its
products faster than most others: for the last 40 years a dollar has
bought hardware with twice as much computation, storage. and communication
every 18 to 24 months, offering a 100-fold gain every decade (Patterson and Hennessy,
1994, p. 21). This rate will continue at least for the next decade (see Chapter
1, Figure 1.1). Better hardware in turn makes it feasible to create
software for new applications: electronic and mechanical design.
climate mapping, digital libraries, desktop publishing, video editing, and telemedicine
are just a few examples. Such applications are often brought to market by new companies
such as Microsoft and Sun Microsystems, both of which produce revenues of more
than $4 billion per year (Computer Select, 1994) and neither of which existed
15 years ago.
The information technology industry is characterized by
great importance to the economy and society, rapid and continuing change, a 10-
to 15-year cycle from major idea to commercial success, and
successive waves of new companies. In this environment a broad program of
publicly funded research is essential for two reasons:
· First,
industrial efforts cannot replace government investment in basic computing and communications
research. Few companies will invest for a payoff that is 10 years away, and
even a company that does make a discovery may postpone using it. The vitality
of the information technology industry depends heavily on new companies, but
new companies cannot easily afford to do research; furthermore, industry in
general is doing less research now than in the recent past
(Geppert, 1994; Corcoran, 1994). But because today's sales are based
on yesterday's research, investment in innovation must go forward so that the
nation's information industry can continue to thrive.
· Second,
it is hard to predict which new ideas and approaches will succeed. The exact course
of exploratory research cannot be planned in advance, and its progress cannot
be measured precisely in the short term. The purpose of
publicly funded research is to advance knowledge and create new
opportunities that industry can exploit in the medium and long term,
not to determine how the market should develop.
THE HIGH PERFORMANCE
COMPUTING
AND COMMUNICATIONS INITIATIVE
Goals and Emphases
The HPCCI is the current manifestation of the
continuing government research program in information technology, an investment
that has been ongoing for more than 50 years. Although it emphasizes research
in high-performance computing and communications, the HPCCI now has in its budget
nearly all of the federal funding for computing research of any kind. The
wisdom of this arrangement is doubtful.
The HPCCI was initiated to serve several broad goals
(NCO. 1993):
· Extend
U.S. leadership in high-performance computing and networking;
· Disseminate
new technologies to serve the economy, national security, education, health care,
and the environment; and
· Spur
gains in U.S. productivity and industrial competitiveness.
The original plans to achieve these goals called
for creating dramatically faster computers and networks,
stretching their limits with Grand Challenge problems in scientific computing,
setting up supercomputer centers with the machines and experts
needed to attack these challenges, and training people to build and
exploit the new technology. More recently the focus has been shifting toward
broader uses of computing and communications.
High Performance
"High performance"—which involves
bringing more powerful computing and communications technology
to bear on a problem—has enabled advances on several fronts. High-performance
systems, for example. deliver answers sooner for complex problems that need large
amounts of computing. Timely and accurate forecasting of weather, mapping of
oil reservoirs, and imaging of tumors are among the benefits
encompassed by the goals listed above. But "high performance,"
which is broader than supercomputing, is a moving target because of the steady
and rapid gains in the performance/cost ratio. Yesterday's
supercomputer is today's personal computer; today's
leading-edge communications technology will be among tomorrow's mainstream capabilities.
Information technology evolves as new and
valuable applications are found for hardware that gets steadily
more powerful and cheaper. To benefit, users need affordable hardware, but they
also need the software that implements the new
applications. Yet learning how to build software takes many years of
experimentation. If this process starts only when the hardware has already become
cheap, the benefits to users will be delayed by years. Research needs to treat
today's expensive equipment as a time machine, learning how it
will be used when it is cheap and widely available, as it surely
will be tomorrow. Knowing how to use computers for new tasks sooner can help
many industries to become more competitive.
To date, the HPCCI's focus has been mainly on speed, but
speed is not the only measure of high performance. Both
speed, measured today in billions of operations per second or billions of bits
per second, and scale, measured by the number of millions of users served, are
important research issues. However, for the nation's information
infrastructure, scale now seems more difficult to achieve.
Information technology can be thought of as a tent, with the height of the center
pole as speed and the breadth of the base as scale. Widening the tent to allow
more work on scale without decreasing the work on speed
requires more cloth; with the same resources, widening
the tent would sacrifice research on speed for research on
scale. This report recommends ways to reallocate funds within the
HPCCI so as to accommodate greater emphasis on scale.
Accomplishments to Date
The HPCCI has focused mainly on parallel
computing, networking, and development of human resources. Building on progress
in research begun before the HPCCI, work and accomplishments to
date reveal two key trends: better computing and computational infrastructure and
increasing researcher-developer-user synergy.
Despite the difficulty of measuring impact at this early
stage, it is the committee's judgment that the HPCCI has been generally
successful so far. That assessment is necessarily qualitative and experiential
now. Because the HPCCI is only 3 years old, results that can be measured in
dollars should not be expected before the next decade.
The HPCCI has contributed substantially to the
development, deployment, and understanding of computing and
communications facilities and capabilities as infrastructure. It has helped transform
understanding of how to share resources and information, generating proofs of
concept and understanding that are of value not only to the
scientific research community but also to the economy and
society at large.
In parallel computing the fundamental challenge
is not building the machines, but learning how to program them.
Pioneering users and their software developers must be motivated by machines
that are good enough to reward success with significant speedups.' For this
reason. a great deal of money and effort have had to be spent to
obtain parallel machines with the potential to run much faster
than existing supercomputers. From the base built by the HPCCI, much has been learned
about parallel computing.
The HPCCI has fostered productive interactions
among the researchers and developers involved in creating
high-performance computing and communications technology and researchers who
use the technology. Building on the varying perspectives of the three groups,
complex problems are being solved in unique ways. In particular,
the I--IPCCI has funded cross-disciplinary teams associated
with the Grand Challenge projects to solve complex computational problems and produce
essential new software for the new parallel systems.
More specifically, the HPCCI has:
· Increased
the nation's stock of expertise by educating new students and attracting new researchers;
· Made
parallel computing widely accepted as the practical route to achieving high-performance
computing;
· Demonstrated
the feasibility of and initiated deployment of parallel databases;
· Driven
progress on Grand Challenge problems in disciplines such as cosmology,
molecular biology, chemistry, and materials science. Parallel
computation has enhanced the ability to attack problems of great complexity in
science and engineering;
· Developed
new modes of analyzing and visualizing complex data sets in the earth sciences,
medicine, molecular biology, and engineering, including creating virtual
reality technologies. Many supercomputer graphic techniques of the
1980s are now available on desktop graphics workstations;
· Through
the gigabit network testbeds associated with the National Research and
Education Network component, demonstrated the intimate link between computing
and communications systems;
· Built
advanced networks that are the backbone of the Internet and the prototypes for
its further evolution into the basis for a broader information infrastructure;
· Deployed
a high-speed backbone that has kept up with the yearly doubling of the size of the
Internet, and organized the impending transition of this backbone away from
government funding; and
· Created
the Mosaic browser for the World Wide Web, the first new major application in many
years that promises to greatly increase access to the resources available on
the Internet. This was an entirely unexpected result.
Evolution
A large-scale, integrated information
infrastructure designed to serve the entire nation is becoming
a high priority for government and industry as well as a source of challenges
for research. Complex systems with millions of users pose many
problems: performance, management, security, interoperability,
compatible evolution of components, mobility, and reliability are only a few.
Today's technology can solve these problems for systems with a few thousand
users at most; to do so for millions or hundreds of millions of users
is far beyond the current state of the art.' Providing users
with high-bandwidth connections is itself a major problem, but it is only the beginning.
There is a wide gap between enabling a connection and providing a rich array of
useful and dependable services.
Because the HPCCI has become the rubric under which
virtually all of the nation's research in information technology is
conducted, it is not surprising that its focus has been changing in response to
past successes, new opportunities, and evolving societal needs. The recently
added Information Infrastructure Technology and Applications (IITA) program,
broadly construed, addresses many of the problems just mentioned; it
is already the largest component of the HPCCI,5 and
its continued evolution should be encouraged.
But with the policy focus—in the government, the press,
and in most of the agencies—centered on information infrastructure,6
high-performance computing seems to have been downplayed. The
committee emphasizes the importance of retaining the HPCCI's momentum at just
the time when its potential to support improvement in the nation's information
infrastructure is most needed.
Organization
Several federal agencies participate in the
HPCCI, most notably the National Science Foundation (NSF), the Advanced
Research Projects Agency (ARPA). the National Aeronautics and Space
Administration (NASA), and the Department of Energy (DOE) (see Appendix A for a
full list). Because of its successes the HPCCI has become a
model for multiagency collaboration and for the "virtual
agency" concept advanced through the National Performance Review (Gore, 1993). Each
participating agency retains responsibility for its own part of the program,
but the agencies work together in joint funding of projects, such
as the federal portions of the Internet; joint reviews of
grants and contracts, such as the NSF-ARPA-NASA digital library initiative;
joint testbeds; and
consortia, such as the consortium for Advanced Modeling of
Regional Air Quality that joins six federal agencies with several state and
local governments.
The HPCCI supports a diverse set of contractors at
universities, companies, and national laboratories throughout the
country. It provides project funding in varying amounts through contracts.
grants, and cooperative agreements awarded according to diverse methods. This
diversity is healthy because it allows many views to compete, resulting in a
broad research program that ensures a continuing flow of
advances in information technology.
Some have argued for a more centrally managed program,
with thorough planning, precise milestones, and presumably
no wasted effort. Tighter management would cost more in bureaucracy and
turf wars, but the essential question is whether it would produce better or
worse results for the money spent. The committee believes that because of the
long time scale of research, diversity is essential for success. No
one person or organization is either smart or lucky enough to plan the best program,
no single approach is best, and success often comes in unanticipated ways.
Because it is a national research program and because of the many
different but interdependent underlying technologies, the HPCCI is
necessarily and properly far more diverse than a focused effort such as the
Apollo moon landing program or a commercial product development program.
In contrast to central management, coordination
enhances the benefits of diversity by helping to prevent
unintended duplication. redundancy, and missed opportunities. The HPCCI's
National Coordination Office (NCO) serves this purpose. aiding
interagency cooperation and acting as liaison for the initiative
to the Congress. other levels of government, universities, industry, and the
public. Its efforts are reflected in its impressive FY 1994 and FY
1995 "Blue Books- describing the program's
activities and spending.' Strengthening the NCO and appointing an advisory
committee, as recommended in the committee's interim report (CSTB.
1994c), would facilitate regular infusions of ideas and advice
from industry and academia and enable better communication of the HPCCI's goals
and accomplishments to its many constituents. This committee should consist of
a group of recognized experts that is balanced between academia and industry
and balanced with respect to application areas and the
core technologies underlying the HPCCI.
Budget
Because it grew from earlier programs. a
significant portion of the HPCCI budget is not new money.
The budget grew from a base of $490 million in preexisting funding in FY 1992
to the $1.1 billion requested for FY 1995.8 Each year agencies have
added to the base by moving budgets for existing programs into the
HPCCI and by reprogramming existing funds to support the HPCCI. Congress
has also added funding each year to start new activities or expand old ones.
The result is that much of the $1.1 billion requested for
FY 1995 is money that was already being spent on computing
and communications in FY 1992. The request has three elements: (1) funds
for activities that predate the HPCCI and were in the FY 1992 base budget, (2)
funds for activities that have since been designated as part of the
HPCCI, and (3) new funds for new activities or for growth. Although dissecting
the budget in this way would shed light on the program, the committee
was unable to do so because each participating agency treats the numbers
differently.
It appears that the FY 1995 request breaks down roughly
as one-third for applications, one-third to advance the
essential underlying computing and communications technologies, one-quarter
for computing and communications infrastructure, and small amounts for
education and electronics (see Appendix C).
THE FUTURE OF THE HPCCI:
RECOMMENDATIONS
The committee believes that strong public support
for a broadly based research program in information technology is
vital to maintaining U.S. leadership in information technology. Incorporating
this view of the importance and success of the government's investment in
research. the 13 recommendations that follow address five areas:
general research program, high-performance computing.
networking and information infrastructure. the supercomputer centers and the
Grand Challenge projects, and program coordination and
management. Within each area the recommendations are
presented in priority order.
General Recommendations
1. Continue
to support research in information technology. Ensure that the major funding
agencies, especially the National Science Foundation and the Advanced Research Projects
Agency, have strong programs for computing and communications research that are
independent of any special initiatives.
The government investment in computing research has
yielded significant returns. Ongoing investment, at least as high
as the current dollar level, is critical both to U.S.
leadership and to ongoing innovation in information technology.
Today the HPCCI supports nearly all of this research, an
arrangement that is both misleading and dangerous: misleading because much important
computing research addresses areas other than high performance (even though it
may legitimately fit under the new IITA component of the HPCCI), and dangerous
because reduced funding for the HPCCI could cripple all of
computing research. The "war on cancer" did not support
all of biomedical research, and neither should the HPCCI or any future
initiative on national infrastructure subsume all of computing
research.
2. Continue
the HPCCI, maintaining today's increased emphasis on the research challenges
posed by the nation's evolving information infrastructure. The
new Information Infrastructure Technology and Applications
program of the HPCCI focuses on information infrastructure topics, which are
also addressed in the initiative's other four components. The committee
supports this continued evolution. which will lead to tangible returns on
existing and future investments in basic hardware, networking.
and software technologies.
High-Performance Computing
3. Continue
funding a strong experimental research program in software and algorithms for
parallel computing machines. Today a crucial obstacle to
widespread use of parallel computing is the lack of advanced
software and algorithms. Emphasis should be given to research
on developing and building usable applications-oriented software systems for
parallel computers. Avoid funding the transfer
("porting") of existing commercial applications to new parallel
computing machines unless there is a specific research need.
4. Stop
direct HPCCI funding for development of commercial hardware by computer vendors
and for "industrial stimulus" purchases of hardware. Maintain HPCCI
support for precompetitive research in computer architecture; this work should
be done in universities or in university-industry
collaborations and should be driven by the needs of system and application
software. HPCCI funding for stimulus purchase of
large-scale machines has been reduced, as has the funding of hardware
development by vendors. The committee supports these changes,
which should continue except when a mission need demands the development of nonstandard
hardware.
Public research is best done in universities. Not
only are academic organizations free to think about longer-term
issues, but they also stimulate technology transfer through publication and placement
of graduates. The national experience supports Vannevar Bush's basic tenet:
publicly funded research carried out in universities produces a
diversity of excellent ideas, trained people, research results,
and technologies that can be commercially exploited (OSRD and Bush, 1945).
5. Treat
development of a teraflop computer as a research direction rather than a destination.
The goal of developing teraflop capability has served a
valuable purpose in stimulating work on large-scale
parallelism, but further investment in raw scalability is inappropriate except
as a focus for precompetitive, academic research. Industrial
development of parallel computers will balance the low cost of individual,
mass-produced computing devices against the higher cost of communicating
between them in a variety of interesting ways. In the near future a teraflop
parallel machine will be built when some agencies' mission
requirements correspond to a sufficiently economical commercial
offering. Continued progress will surely lead to machines even larger than a
teraflop.
Networking and Information
Infrastructure
New ideas are needed to meet the new challenges
underlying development of the nation's information infrastructure.
The HPCCI can contribute most by focusing on the underlying research issues.
This shift has already begun, and it should continue.
This evolution of the research agenda, which would support
improvement of the nation's information infrastructure,
is partly under way: in the FY
1995 Implementation Plan (NCO, 1994, p.
15), over one-quarter of the NSF and ARPA HPCCI funding is focused on the IITA
component. and activities in other components have also evolved
consistent with these concerns. The committee supports this
increased emphasis.
6. Increase
the HPCCI focus on communications and networking research, especially on
the challenges inherent in scale and physical distribution. An
integrated information infrastructure that fully serves the nation's needs
cannot spring full-grown from what we already know. Much research
is needed on difficult problems related to size. evolution, introduction of new
systems, reliability, and interoperability. Much more is
involved than simply deploying large numbers of boxes and wires.
For example, both hardware and software systems must work efficiently
to handle scheduling; bandwidth optimization for transmission of a range of
data formats, including real-time audio and video data; protocol and format
conversion; security; and many other requirements.
7. Develop
a research program to address the research challenges underlying our ability
to build very large, reliable, high-performance, distributed information
systems based on the existing HPCCI foundation. Although
a comprehensive vision of the research needed for advancing
the nation's information infrastructure has not yet been developed, three key
areas for research are scalability, physical distribution, and
interoperative applications.
8. Ensure
that research programs focusing on the National Challenges contribute to the
development of information infrastructure technologies as well as to the
development of new applications and paradigms. This
dual emphasis contrasts with the narrower focus on scientific
results that has driven work on the Grand Challenges.
Supercomputer Centers and
Grand Challenge Program
The NSF supercomputer centers have played a
major role in establishing parallel computing as a full partner with the prior
paradigms of scalar and vector computing by providing access to
state-of-the-art computing facilities. NSF should continue
to take a broad view of the centers' mission of providing
access to high-performance computing and communications resources, including
participating in research needed to improve software for parallel machines and
to advance the nation's information infrastructure.
The committee recognizes that advanced computation is an
important tool for scientists and and engineers and that support for adequate
computer access must be a part of the NSF research program in all disciplines.
The committee did not consider the appropriate overall funding level for the
centers. However, the committee believes that NSF should move to a model
similar to that used by NASA and DOE for funding general access to
computing. The committee prefers NASA's and DOE's approach to
funding supercomputer centers, where HPCCI funds are used only to support the exploration
and use of new computing architectures, while non-HPCCI funds are used to
support general access.
9. The
mission of the National Science Foundation supercomputer centers remains important,
but the NSF should continue to evaluate new directions, alternative funding mechanisms,
new administrative structures, and the overall program level of the centers.
NSF could continue funding of the centers at the current level
or alter that level, but it should continue using HPCCI funds
to support applications that contribute to the evolution of the underlying
computing and communications technologies, while support for general access by application
scientists to maturing architectures should come increasingly from non-HPCCI funds.
10. The
Grand Challenge program is an innovative approach to creating interdisciplinary
and multi-institutional scientific research teams; however, continued use of HPCCI
funds is appropriate only when the research contributes significantly to the development
of new high-performance computing and communications hardware or software.
Grand Challenge projects funded under the HPCCI should be evaluated on the
basis of their contributions both to high-performance computing
and communications technologies and to the application
area. Completion of the Grand Challenge projects will provide
valuable insights and demonstrate the capabilities of new
high-performance architectures in some important applications. It
will also foster better collaboration between computer scientists and
computational scientists. The committee notes that a large
share of HPCCI funding for the Grand Challenges currently comes
from the scientific disciplines involved. However, the overall funding seems to
come entirely from HPCCI-labeled funds. For the same
reasons outlined in Recommendation 9, the committee sees this
commingled support as unhealthy in the long run and urges a transition to greater
reliance on scientific disciplinary funding using non-HPCCI funds.
Coordination and Program Management
11. Strengthen the HPCCI National Coordination
Office while retaining the cooperative structure of the
HPCCI and increasing the opportunity for external input. Immediately
appoint the congressionally mandated advisory committee intended to provide broad-based,
active input to the HPCCI, or provide an effective alternative. Appoint an individual
to be a full-time coordinator, program spokesperson, and advocate for the
HPCCI.
In making this recommendation, the committee strongly
endorses the role of the current NCO as supporting the mission agencies rather
than directing them. The committee believes it vital that
the separate agencies retain direction of their HPCCI funds. The value of
interagency cooperation outweighs any benefits that might be gained
through more centralized management.
Diverse management systems for research should be
welcomed, and micromanagement should be avoided. In the
past, choosing good program officers and giving them freedom to operate independently
have yielded good value, and the committee believes it will continue to do so.
Furthermore,
independence will encourage diversity in the research program, thus increasing opportunities
for unexpected discoveries, encouraging a broader attack on problems. and
ensuring fewer missed opportunities.
12. Place
projects in the HPCCI only if they match well to its objectives. Federal research
funding agencies should promptly document the extent to which HPCCI funding is supporting
important long-term research areas whose future funding should be independent
of the future of the
HPCCI.
A number of preexisting agency programs have
entered the HPCCI, with two effects: the HPCCI's budget appears to
grow faster than the real growth of investment in high-performance computing
and communications research, and important programs such as basic research in computing
within NSF and ARPA may be in jeopardy should the FIPCCI end.
13. Base
mission agency computer procurements on mission needs only, and encourage making
equipment procurement decisions at the lowest practical management level. This
recommendation applies equally to government agencies and to government
contractors. It has generally been best for an agency to specify the results it
wants and to leave the choice of specific equipment to the contractor
or local laboratory management.
NOTES
1.
See U.S. DOC (1994); the Department of Commerce utilizes
data from the U.S. Bureau of the Census series. the Annual Survey of
Manufactures. It places the value of shipments for the information
technology industry at $421 billion for 1993. This number omits revenue from
equipment rentals, fees for after-sale service, and mark-ups in the product
distribution channel. It also excludes office equipment in total. It includes
computers, storage devices, terminals and peripherals; packaged software;
computer program manufacturing, data processing, information services,
facilities management, and other services; and telecommunications equipment and
services.
See also CBEMA (1994); CBEMA values the worldwide 1993
revenue of the U.S. information technology industry at $602 billion. In addition to
including office equipment, it shows larger revenues for information technology
hardware
and telecommunications equipment than does the Department of Commerce.
2.
Microcomputers (personal computers) are defined as computers
with a list price of $1,000 to $14,999; see CBEMA (1994), pp. 60-61. Forrester
Research Inc. (1994, pp. 2-3) estimates the share of households with PCs at
about 20
percent, based on its survey of households and Bureau of Census data.
Forrester's model accounts for retirements of older PCs and for households with
multiple PCs. This is a lower estimate than the Software Publishing
Association's widely cited 30 percent share. By definition, the microcomputer
statistics exclude small computers and other general-purpose and specialized
devices that also make use of microprocessors and would be counted in a more comprehensive
measurement of information technology.
3.
Earlier experience with three isolated computers,
"Illiac 4" (built at the University of Illinois) and
"C.mmp" and -Cm*" (both built at Carnegie Mellon University),
bears out this point.
4.
Of course, systems specialized for a single application or
for homogenous technology, such as telephony, serve millions of users, but what
is now envisioned is more complex and heterogenous, involving integration of
multiple services and systems.
5.
The other four programs of the HPCCI are Advanced Software
Technology and Algorithms, Basic Research and Human Resources, High-Performance
Computing Systems, and the National Research and Education Network.
6.
Notably, references to the computing portion of the HPCCI
have been overshadowed recently by the ubiquity of speeches and documents devoted
to the notion of a national information infrastructure (Nil). The NII has also
been featured
in the titles of the 1994 and 1995 Blue Books.
7.
Each year beginning in 1991 the director of the Office of
Science and Technology Policy submits a report on the TIPCCI to
accompany the president's budget. The FY 1992, FY 1993, and FY 1994 hooks were produced
by the now-defunct Federal Coordinating Council for Science, Engineering, and
Technology; the FY 1995 report was produced by the NCO (acting for the Committee on
Information and Communications). The report describes prior accomplishments and the future
funding and activities for the coming fiscal year. These reports have
collectively become known as "Blue Books" after the color of their
cover.
8.
NCO (1994), p. 15. Note that figures represent the President's requested budget authority
for FY 1995. Actual appropriated levels were not available at press time. Because the
HPCCI is synthesized as a cross-cutting multiagency initiative, there is no
separate and identifiable "HPCCI appropriation."
COMMITTEE TO
STUDY
HIGH PERFORMANCE COMPUTING AND COMMUNICATIONS:
STATUS OF A MAJOR INITIATIVE
FREDERICK
P. BROOKS, JR., University of North Carolina at Chapel Hill, Co-chair IVAN
E. SUTHERLAND, Sun Microsystems Laboratories, Co-chair ERICH BLOCH, Council on
Competitiveness
DEBORAH ESTRrN, University of Southern California/Information
Sciences Institute JOHN HENNESSY, Stanford University
BUTLER W. LAMPSON, Digital Equipment Corporation EDWARD
D. LAZOWSKA. University of Washington
WILLIAM A. LESTER. JR., University of California at
Berkeley JANE PRESTON, Telemedical Interactive Consultative
Services Inc. W. DAVID SINCOSKIE. Bell Communications Research
Inc.
LARRY
SMARR. National Center for Supercomputing Applications/University of
Illinois at Urbana-Champaign
JOSEPH
F. TRAUB, Columbia University
Staff
MARJORY S. BLUMENTHAL, Director JAMES
E. MALLORY, Staff Officer LESLIE M. WADE. Project Assistant
COMPUTER SCIENCE AND TELECOMMUNICATIONS BOARD
WILLIAM WULF, University of Virginia, Chair FRANCES ALLEN, IBM T.J. Watson Research Center JEFF DOZIER, University of California at Santa Barbara DAVID J. FARBER, University of Pennsylvania HENRY FUCHS, University of North Carolina CHARLES M. GESCHKE, Adobe Systems Inc. JAMES GRAY, San Francisco, California BARBARA J. GROSZ,
Harvard University DEBORAH A. JOSEPH,
University of Wisconsin RICHARD M. KARP, University
of California at Berkeley BUTLER W. LAMPSON,
Digital Equipment Corporation BARBARA H. LISKOV,
Massachusetts Institute of Technology
JOHN MAJOR, Motorola Inc.
ROBERT
L. MARTIN, AT&T Network Systems
DAVID
G. MESSERSCHMITT. University of California at Berkeley
WILLIAM
H. PRESS, Harvard University CHARLES L.
SEITZ, Myricom Inc.
EDWARD SHORTLIFFE, Stanford University School of Medicine
CASMIR S. SKRZYPCZAK, NYNEX Corporation LESLIE L. VADASZ, Intel Corporation
MARJORY S. BLUMENTHAL, Director LOUISE
A. ARNHEIM, Senior Staff Officer HERBERT S.
L1N, Senior Staff Officer JAMES E.
MALLORY, Staff Officer RENEE A.
HAWKINS, Staff Associate JOHN M.
GODFREY, Research Associate GLORIA P.
BEMAH, Administrative Assistant LESLIE
M. WADE, Project Assistant