[House Hearing, 110 Congress]
[From the U.S. Government Publishing Office]
OVERSIGHT OF THE NETWORKING
AND INFORMATION TECHNOLOGY RESEARCH
AND DEVELOPMENT (NITRD) PROGRAM
=======================================================================
HEARING
BEFORE THE
COMMITTEE ON SCIENCE AND TECHNOLOGY
HOUSE OF REPRESENTATIVES
ONE HUNDRED TENTH CONGRESS
SECOND SESSION
__________
JULY 31, 2008
__________
Serial No. 110-119
__________
Printed for the use of the Committee on Science and Technology
Available via the World Wide Web: http://www.science.house.gov
______
U.S. GOVERNMENT PRINTING OFFICE
43-688 PDF WASHINGTON DC: 2008
---------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office Internet: bookstore.gpo.gov Phone: toll free (866)512-1800
DC area (202)512-1800 Fax: (202) 512-2250 Mail Stop SSOP,
Washington, DC 20402-0001
COMMITTEE ON SCIENCE AND TECHNOLOGY
HON. BART GORDON, Tennessee, Chairman
JERRY F. COSTELLO, Illinois RALPH M. HALL, Texas
EDDIE BERNICE JOHNSON, Texas F. JAMES SENSENBRENNER JR.,
LYNN C. WOOLSEY, California Wisconsin
MARK UDALL, Colorado LAMAR S. SMITH, Texas
DAVID WU, Oregon DANA ROHRABACHER, California
BRIAN BAIRD, Washington ROSCOE G. BARTLETT, Maryland
BRAD MILLER, North Carolina VERNON J. EHLERS, Michigan
DANIEL LIPINSKI, Illinois FRANK D. LUCAS, Oklahoma
NICK LAMPSON, Texas JUDY BIGGERT, Illinois
GABRIELLE GIFFORDS, Arizona W. TODD AKIN, Missouri
JERRY MCNERNEY, California TOM FEENEY, Florida
LAURA RICHARDSON, California RANDY NEUGEBAUER, Texas
PAUL KANJORSKI, Pennsylvania BOB INGLIS, South Carolina
STEVEN R. ROTHMAN, New Jersey DAVID G. REICHERT, Washington
JIM MATHESON, Utah MICHAEL T. MCCAUL, Texas
MIKE ROSS, Arkansas MARIO DIAZ-BALART, Florida
BEN CHANDLER, Kentucky PHIL GINGREY, Georgia
RUSS CARNAHAN, Missouri BRIAN P. BILBRAY, California
CHARLIE MELANCON, Louisiana ADRIAN SMITH, Nebraska
BARON P. HILL, Indiana PAUL C. BROUN, Georgia
HARRY E. MITCHELL, Arizona VACANCY
CHARLES A. WILSON, Ohio
ANDRE CARSON, Indiana
C O N T E N T S
July 31, 2008
Page
Witness List..................................................... 2
Hearing Charter.................................................. 3
Opening Statements
Statement by Representative Bart Gordon, Chairman, Committee on
Science and Technology, U.S. House of Representatives.......... 7
Written Statement............................................ 8
Statement by Representative Ralph M. Hall, Minority Ranking
Member, Committee on Science and Technology, U.S. House of
Representatives................................................ 9
Written Statement............................................ 9
Prepared Statement by Representative Eddie Bernice Johnson,
Member, Committee on Science and Technology, U.S. House of
Representatives................................................ 10
Prepared Statement by Representative Harry E. Mitchell, Member,
Committee on Science and Technology, U.S. House of
Representatives................................................ 11
Witnesses:
Dr. Christopher L. Greer, Director, National Coordination Office
for Networking and Information Technology Research and
Development (NCO/NITRD)
Oral Statement............................................... 11
Written Statement............................................ 13
Biography.................................................... 26
Dr. Daniel A. Reed, Director, Scalable and Multicore Computing,
Microsoft Corporation
Oral Statement............................................... 26
Written Statement............................................ 28
Biography.................................................... 35
Dr. Craig A. Stewart, Chair, Coalition for Academic Scientific
Computing; Associate Dean, Research Technologies, Indiana
University
Oral Statement............................................... 36
Written Statement............................................ 37
Biography.................................................... 41
Mr. Don C. Winter, Vice President, Engineering and Information
Technology, Phantom Works, the Boeing Company
Oral Statement............................................... 42
Written Statement............................................ 44
Biography.................................................... 48
Discussion....................................................... 48
Appendix: Answers to Post-Hearing Questions
Dr. Christopher L. Greer, Director, National Coordination Office
for Networking and Information Technology Research and
Development (NCO/NITRD)........................................ 58
Dr. Daniel A. Reed, Director, Scalable and Multicore Computing,
Microsoft Corporation.......................................... 64
Dr. Craig A. Stewart, Chair, Coalition for Academic Scientific
Computing; Associate Dean, Research Technologies, Indiana
University..................................................... 67
Mr. Don C. Winter, Vice President, Engineering and Information
Technology, Phantom Works, the Boeing Company.................. 70
OVERSIGHT OF THE NETWORKING AND INFORMATION TECHNOLOGY RESEARCH AND
DEVELOPMENT (NITRD) PROGRAM
----------
THURSDAY, JULY 31, 2008
House of Representatives,
Committee on Science and Technology,
Washington, DC.
The Committee met, pursuant to call, at 10:03 a.m., in Room
2318 of the Rayburn House Office Building, Hon. Bart Gordon
[Chairman of the Committee] presiding.
hearing charter
COMMITTEE ON SCIENCE AND TECHNOLOGY
U.S. HOUSE OF REPRESENTATIVES
Oversight of the Networking
and Information Technology Research
Development (NITRD) Program
thursday, july 31, 2008
10:00 a.m.-12:00 p.m.
2318 rayburn house office building
1. Purpose
On Thursday, July 31, 2008, the Committee on Science and Technology
will hold an oversight hearing to review the multi-agency, coordinated
Networking and Information Technology Research and Development (NITRD)
program. The hearing will examine the current program in light of the
recent assessment of the President's Council of Advisors on Science and
Technology (PCAST) and explore whether additional legislative
adjustments to the program are needed.
2. Witnesses
Dr. Chris L. Greer, Director, National Coordination Office for
Networking and Information Technology Research and Development (NCO/
NITRD).
The NCO/NITRD provides staff support for the subcommittees and
working groups of the National Science and Technology Council that are
responsible for planning and coordinating the NITRD program and serves
as the interface with the public for the NITRD program.
Dr. Daniel A. Reed, Director of Scalable and Multicore Computing,
Microsoft.
Dr. Reed is a member of PCAST and of the PCAST committee that
carried out the recent assessment of the NITRD program. He previously
served as a member of the President's Information Technology Advisory
Committee.
Dr. Craig Stewart, Associate Dean, Research Technologies, Indiana
University, and representing the Coalition for Academic Scientific
Computation (CASC).
CASC members are academic and government computer centers that
support computational research in science and engineering and that are
involved in applications requiring high-performance computers and
networks and advanced software development.
Mr. Don C. Winter, Vice President--Engineering and Information
Technology, Phantom Works, the Boeing Company.
Mr. Winter has been involved in a planning effort with others from
industry and academia to develop a research agenda and roadmap in the
area of cyber-physical systems, which is one of the key research areas
the PCAST assessment calls out for increased funding under the NITRD
program.
3. Overarching Questions
Do the objectives of the NITRD program address the
most important information technology R&D issues? Are the R&D
objectives prioritized and are the resources allocated
appropriately to achieve the objectives?
Are there significant research opportunities that the
NITRD program is not pursuing?
Is the overall funding level for the NITRD program
adequate for maintaining U.S. leadership in this important
technology field?
Are any changes needed to the planning, coordination,
and prioritization mechanisms of the NITRD program in order to
make them function more effectively?
Does the research community--both academe and
industry--have a voice in influencing the research priorities
under the NITRD program? Are improvements needed in the
external advisory process for the NITRD program?
Do the recommendations of the recent PCAST assessment
of the NITRD program encompass all of the key issues necessary
to make the NITRD program more effective and relevant to
research needs and opportunities in information technology?
4. Background
NITRD Program
The High-Performance Computing Act of 1991 (P.L. 102-194), which
the Science and Technology Committee was instrumental in enacting,
authorized a multi-agency research program, called the High Performance
Computing and Communications program, to accelerate progress in the
advancement of computing and networking technologies and to support
leading edge computational research in a range of science and
engineering fields. The name of the program has evolved to the
Networking and Information Technology Research and Development (NITRD)
program. The statute established a set of mechanisms and procedures to
provide for the interagency planning, coordination, and budgeting of
the research and development activities carried out under the program.
For FY 2009, 13 federal agencies will contribute funding to the
NITRD program and additional agencies that do not contribute funding
participate in planning activities. The FY 2009 budget request for the
NITRD program is $3.548 billion, an increase of $0.207 billion or
approximately six percent, over the FY 2008 level of $3.341 billion. A
summary of the major research components of the program and funding
levels by major component and by agency is available at: http://
www.nitrd.gov/pubs/2009supplement/index.htm
Assessment of NITRD by the President's Council of Advisors on Science
and Technology (PCAST)
P.L. 102-194 provided for an external advisory committee for the
NITRD program. A subsequent executive order created the President's
Information Technology Advisory Committee (PITAC). The current
Administration allowed that committee to expire and in its place
assigned the advisory function for the NITRD program to PCAST. Last
August PCAST completed an assessment of the NITRD program and issued a
report, ``Leadership Under Challenge: Information Technology R&D in a
Competitive World'' [http://www.nitrd.gov/pcast/reports/PCAST-NIT-
FINAL.pdf].
The PCAST report includes several findings and recommendations
related to the research content of the program, as well as suggestions
for improving the program's planning, prioritization and coordination.
The recommendations from the PCAST report include:
Federal agencies should rebalance their NITRD funding
portfolios by increasing support for important problems that
require larger-scale, longer-term, multi-disciplinary R&D and
increasing emphasis on innovative and therefore higher-risk but
potentially higher-payoff explorations.
As new funding becomes available for the NITRD
program, disproportionately larger increases should go for:
research on NIT systems connected with the physical
world (which are also called embedded, engineered, or
cyber-physical systems);
software R&D;
a national strategy and implementation plan to
assure the long-term preservation, stewardship, and
widespread availability of data important to science
and technology; and
networking R&D, including upgrading the Internet and
R&D in mobile4networking technologies.
The NITRD agencies should:
develop, maintain, and implement a strategic plan
for the NITRD program;
conduct periodic assessments of the major components
of the NITRD program and restructure the program when
warranted;
develop, maintain, and implement public R&D plans or
roadmaps for key technical areas that require long-term
interagency coordination and engagement; and
develop a set of metrics and other indicators of
progress for the NITRD program, including an estimate
of investments in basic and applied research, and use
them to assess NITRD program progress.
The NITRD National Coordination Office should support
the development, maintenance, and implementation of the NITRD
strategic plan and R&D plans for key technical areas; and it
should be more proactive in communicating with outside groups.
Cyber-Physical Systems
The top recommendation of the PCAST report for new research
investments in the NITRD program is in the area of computer-driven
systems connected with the physical world--also called embedded,
engineered, or cyber-physical systems (CPS). CPS are connected to the
physical world through sensors and actuators to perform crucial
monitoring and control functions. Such systems would include the air-
traffic-control system, the power-grid, water-supply systems, and
industrial process control systems. On a more individual level, they
are found in automobiles and home health care devices.
Examples of CPS are already in widespread use but growing demand
for new capabilities and applications will require significant
technical advances. Such systems can be difficult and costly to design,
build, test, and maintain. They often involve the intricate integration
of myriad networked software and hardware components, including
multiple subsystems. In monitoring and controlling the functioning of
complex, fast-acting physical systems (such as medical devices, weapons
systems, manufacturing processes, and power-distribution facilities),
they must operate reliably in real time under strict constraints on
computing, memory, power, speed, weight, and cost. Moreover, most uses
of cyber-physical systems are safety-critical: they must continue to
function even when under attack or stress.
There is evidence that CPS will be an area of international
economic competition. For example, the European Union's Advanced
Research and Technology for Embedded Intelligence and Systems (ARTEMIS)
program, funded by a public-private investment of 5.4 billion euros
(over $7 billion in mid-2007 dollars) between 2007 and 2013, is
pursuing R&D to achieve ``world leadership in intelligent electronic
systems'' by 2016.
Recent Amendments to P.L. 102-194 [included in COMPETES Act]
In 1999, the PITAC released an assessment of the NITRD program
(``Information Technology Research: Investing in Our Future'') that
found the research sponsored to be migrating too much toward support
for near-term, mission focused objectives; that found a growing gap
emerging between the power of high-performance computers available to
support agency mission requirements versus support for the general
academic research community; and that found the total federal
information technology investment inadequate. In response to that
report, the Committee developed legislation that passed the House in
similar form in the 108th (H.R. 4218) and 109th (H.R. 28) Congresses,
but failed to be picked up in the Senate. It was finally incorporated
in the COMPETES Act (section 7024(a) ) in this Congress.
The COMPETES Act amends the 1991 Act in several ways:
Program Planning. Specifies that the external advisory
committee for the program, which must be re-constituted as a
separate stand-alone committee, must carry out biennial reviews
of the funding, content and management of the interagency R&D
program and report its findings to Congress. Also, the annual
report on the program prepared by the OSTP Director must now
describe how the program has been modified in response to
advisory Committee's recommendations.
High-End Computing. Requires OSTP to develop and maintain a
roadmap for developing and deploying very high-performance
computing (high-end) systems necessary to ensure that the U.S.
research community has sustained access to the most capable
computing systems.
Large Scale Applications. Clarifies that Grand Challenge
problems supported under the interagency program are intended
to involve multi-disciplinary teams of researchers working on
science and engineering problems that demand the most capable
high performance computing and networking resources. Consistent
with this requirement, the language also specifies that
provision for access to high performance computing systems
includes technical support to users of these systems.
5. Witness Questions
Dr. Greer was asked to provide an overview of the current planning
and coordination mechanisms of the NITRD program, along with any
recommendations on how to improve their effectiveness; a description of
any actions by the NITRD agencies that have been taken, or that are in
the planning stages, in response to the recommendations of the PCAST
report; a description of the role of the National Coordination Office
in supporting the activities of the NITRD program; and his response to
the findings and recommendations of the PCAST report related to the
functioning of the NCO.
The other witnesses were asked to review and comment on the
findings and recommendations contained in the PCAST report regarding
both the administration and planning for the NITRD program and also the
research priorities that the program should address. They were asked
for their views on the merit of these recommendations and on what they
see as the key steps to take that would strengthen the NITRD program,
including any issues not addressed by the PCAST report.
Mr. Winter was particularly asked to provide his views on the PCAST
recommendation related to the need for the NITRD program to place
greater emphasis on research on cyber-physical systems.
Chairman Gordon. I want to welcome everyone to this
morning's hearing to review the federal, interagency research
initiative in networking and information technology, known as
the NITRD program.
Information technology is a major driver of economic
growth. It creates high-wage jobs, provides for rapid
communication throughout the world, and provides the tools for
acquiring knowledge and insight from information. Advances in
computing and communications have a broad impact. For example,
information technology helps to make the workplace more
productive, to improve the quality of health care, and to make
government more responsive and accessible to the needs of our
citizens. In short, networking and information technology is an
essential component of U.S. scientific, industrial, and
military competitiveness.
Vigorous long-term research is essential to realize the
potential benefits of information technology. Many of the
technical advances that led to today's computers and the
Internet resulted from federally sponsored research, in
partnership with industry and universities.
The depth and strength of U.S. capabilities in information
technology stem in part from the sustained research and
development program carried out by the federal research
agencies under a program codified by the High-Performance
Computing Act of 1991. The Science and Technology Committee has
a long history of encouragement and support for research on
information technologies and played a prominent role in the
development and passage of the 1991 Act.
The Act created a multi-agency R&D program to accelerate
the development of information technology and to attack
challenging computational science and engineering problems.
Also, it put in place a formal process for planning and
budgeting for the activities carried out under what is now
known as the NITRD program.
The fiscal year 2008 budget for this interagency program is
$3.3 billion. The agencies providing the largest portions of
this funding are the Department of Defense, the National
Science Foundation, the Department of Energy, and the National
Institutes of Health.
I believe the NITRD program has been largely a success. It
has made a substantial contribution to moving computation to an
equal place along side theory and experiment for conducting
research in science and engineering.
Moreover, it has developed the computing and networking
infrastructure needed to support leading edge research and to
drive the technology forward for a range of commercial
applications that benefit society broadly.
The President's Council of Advisors on Science and
Technology, the PCAST, recently carried out an assessment of
the NITRD program. This assessment raises some significant
issues about whether the NITRD program is allocating its
resources in ways to achieve the maximum payoffs. PCAST makes a
series of recommendations that identify research areas needing
additional resources and suggests that the modes of research
support provided by the program are less than optimum.
In particular, PCAST believes that the NITRD program should
provide more of its funding for conducting high-risk/high-
reward research and support more large-scale, interdisciplinary
research projects. It also recommends that the NITRD program
institute a strategic planning process to strengthen priority
setting under the program. I believe that PCAST raises issues
that need to be seriously considered and then addressed, as
appropriate, through the legislative adjustments to the NITRD
authorizing statute. This hearing is the first step in a
process, which the Committee will conduct next year.
To assist us in the development of legislation, this
hearing provides the opportunity to receive the views of expert
witnesses on the findings and recommendations of the PCAST
assessment of the NITRD program. I am also interested in any
comments or suggestions the witnesses may have on other aspects
of the program that are not covered by the PCAST assessment but
would lead to a more effective program.
I want to thank our witnesses for their attendance at this
hearing and look forward to our discussions.
[The prepared statement of Chairman Gordon follows:]
Prepared Statement of Chairman Bart Gordon
I want to welcome everyone to this morning's hearing to review the
federal, interagency research initiative in networking and information
technology, known as the NITRD [``ny-ter D''] program.
Information technology is a major driver of economic growth. It
creates high-wage jobs, provides for rapid communication throughout the
world, and provides the tools for acquiring knowledge and insight from
information.
Advances in computing and communications have broad impact. For
example, information technology helps to make the workplace more
productive, to improve the quality of health care, and to make
government more responsive and accessible to the needs of our citizens.
In short, networking and information technology is an essential
component of U.S. scientific, industrial, and military competitiveness.
Vigorous long-term research is essential for realizing the
potential benefits of information technology. Many of the technical
advances that led to today's computers and the Internet resulted from
federally sponsored research, in partnership with industry and
universities.
The depth and strength of U.S. capabilities in information
technology stem in part from the sustained research and development
program carried out by federal research agencies under a program
codified by the High-Performance Computing Act of 1991. The Science and
Technology Committee has a long history of encouragement and support
for research on information technologies and played a prominent role in
the development and passage of the 1991 Act.
The Act created a multi-agency R&D program to accelerate the
development of information technology and to attack challenging
computational science and engineering problems. Also, it put in place a
formal process for planning and budgeting for the activities carried
out under the NITRD program.
The Fiscal Year 2008 budget for this interagency program is $3.3
billion. The agencies providing the largest portions of this funding
are the Department of Defense, the National Science Foundation, the
Department of Energy, and the National Institutes of Health.
I believe the NITRD program has been largely a success. It has made
a substantial contribution to moving computation to an equal place
along side theory and experiment for conducting research in science and
engineering.
Moreover, it has developed the computing and networking
infrastructure needed to support leading edge research and to drive the
technology forward for a range of commercial applications that benefit
society broadly.
The technical advances that led to today's computing devices and
networks, and the software that drive them, evolved from past research
sponsored by industry and government, often in partnership, and
conducted by industry, universities, and federal labs.
The President's Council of Advisors on Science and Technology--the
PCAST--recently carried out an assessment of the NITRD program. This
assessment raises some significant issues about whether the NITRD
program is allocating its resources in ways to achieve the maximum
payoffs.
PCAST makes a series of recommendations that identify research
areas needing additional resources and suggests that the modes of
research support provided by the program are less than optimum.
In particular, PCAST believes that the NITRD program should provide
more of its funding for conducting high-risk/high-reward research and
support more large scale, interdisciplinary research projects. It also
recommends that the NITRD program institute a strategic planning
process to strengthen priority setting under the program.
I believe that PCAST raises issues that need to be seriously
considered and then addressed, as appropriate, through legislative
adjustments to the NITRD authorizing statute. This hearing is the first
step in a process, which the Committee will conclude next year.
To assist us in the development of legislation, this hearing
provides the opportunity to receive the views of expert witnesses on
the findings and recommendations of the PCAST assessment of the NITRD
program.
I am also interested in any comments or suggestions the witnesses
may have on other aspects of the program that are not covered by the
PCAST assessment, but which could lead to a more effective program.
I want to thank our witnesses for their attendance at this hearing
and look forward to our discussion.
Chairman Gordon. And now I recognize my friend, Mr. Hall,
for his opening statement.
Mr. Hall. Thank you, Chairman Gordon, for scheduling the
oversight hearing of the NITRD Program. Of course, this program
provides the primary mechanism by which the Federal Government
coordinates this nation's more than $3 billion of unclassified
networking information technology research and development
investments, and you are absolutely correct in highlighting the
PCAST report. To boil it down they simply said, ``It is
essential to U.S. economic prosperity, security, and quality of
life.'' So given the ever-increasing amounts of networking and
information technology that affect our everyday lives from
power grid and water purification systems to automotive
improvements and air traffic control equipment to home health
care devices and educational software programs, it is important
that we not only continue to support these R&D efforts but also
make sure that this program is appropriately coordinating with
our classified Cyber Security Initiatives as well. In fact, I
believe that this is of vital importance to our homeland
security and to our economy.
We have before us today an esteemed panel of NIT experts,
and I look forward to hearing their views and how to make an
already exemplary interagency program even better, and I yield
back my time.
[The prepared statement of Mr. Hall follows:]
Prepared Statement of Representative Ralph M. Hall
Thank you Chairman Gordon for scheduling this oversight hearing on
the NITRD program. This program provides the primary mechanism by which
the Federal Government coordinates this nation's more than three
billion dollars of unclassified networking and information technology
(NIT) research and development (R&D) investments.
The United States is the global leader in NIT, and I agree with the
authors of the PCAST Report on this issue when they say that our
continued leadership ``is essential to U.S. economic prosperity,
security, and quality of life.''
Given the ever increasing amounts of networking and information
technology that affect our everyday lives from power grid and water
purification systems to automotive improvements and air traffic control
equipment to home health-care devices and educational software
programs, it is important that we not only continue to support these
R&D efforts but also make sure that this program is appropriately
coordinating with our classified cyber security initiatives as well. In
fact, I believe that this is of vital importance to our homeland
security and to our economy.
We have before us today an esteemed panel of NIT experts, and I
look forward to hearing their views on how to make an already exemplary
interagency program even better.
Chairman Gordon. Thank you, Mr. Hall. At this time we
normally ask that Members who want to submit their opening
statements, but since Mr. Sensenbrenner, because of his new
status, we will allow him to make any statements that he would
like to at this time.
Mr. Sensenbrenner. Mr. Sensenbrenner is a man of few words
unless they are really necessary. Not this morning. I thank the
Chair.
Chairman Gordon. All right. If that is the case, then if
Members would like to submit additional opening statements,
they will be added to the record at this point.
[The prepared statement of Ms. Johnson follows:]
Prepared Statement of Representative Eddie Bernice Johnson
Good morning, Mr. Chairman and Ranking Member.
Today's hearing on oversight of the Networking and Information
Technology Research and Development (NITRD) Program is important.
The Committee on Science and Technology is tasked with the
important activity of seeing that our federal resources are allocated
responsibly.
Investment in information technology is important to our nation,
and it is important to Texas.
Texas Instruments, located in Dallas, has been a leader in
information technology. Our state has been a welcoming place for high-
tech development, as is evidenced by cities like Austin, Dallas and
Houston gaining reputation as high-tech hubs.
The President's Council of Advisors on Science and Technology
recently assessed the NITRD Program, and today's hearing will be
important to determine whether the objectives of the program address
the most important information technology R&D issues facing our nation.
We need to know if this program's research objectives are
prioritized well, and whether currently allocated resources are
appropriate to achieve these objectives.
It is the Committee's distinct honor to have witnesses representing
the NITRD program, as well as the President's Council of Advisors on
this issue, as well as academic and industry representatives who can
offer unique perspectives.
Computer networking has become very sophisticated. Systems of
computerized sensors perform crucial monitoring and control functions.
Such systems include the air-traffic-control system, the power-grid,
water-supply systems, and industrial process control systems. On a more
individual level, they are found in automobiles and home health care
devices.
Our nation must continue to innovate and stay at the leading edge
of this kind of technology. Other nations are currently investing
heavily in this activity and are gaining competitive ground.
Take, for example, the European Union's Advanced Research and
Technology for Embedded Intelligence and Systems (ARTEMIS) program.
This effort is funded by a public-private investment of 5.4 billion
euros (over $7 billion in mid-2007 dollars) between 2007 and 2013, is
pursuing research to achieve ``world leadership in intelligent
electronic systems'' by 2016.
President's Council of Advisors on Science and Technology have
provided guidance on how our NITRD program should move forward.
Principal among their recommendations is the suggestion to
rebalance agency funding portfolios to support more long-term projects,
as well as research that is considered to be high-risk.
The Council also advised that greater proportions of funding should
go toward research in networking information technology systems that
are connected with the physical world; and for mobile networking
technologies.
In addition, the Council urged agencies to implement a strategic
plan for the NITRD program; and they should develop metrics to assess
the progress of investments in research in these areas.
I want to thank the Council for their work to assess the NITRD
program.
I hope that the recommendations, as well as this hearing, will help
the Science Committee continue to be a good steward when it comes to
allocation of funds for computer networking research.
[The prepared statement of Mr. Mitchell follows:]
Prepared Statement of Representative Harry E. Mitchell
Thank you, Mr. Chairman.
Today we will examine current status of the Networking and
Information Technology Research and Development (NITRD) program.
In 1999, an assessment of the NITRD program by the President's
Information Technology Advisory Committee (PITAC) concluded that the
research sponsored by NITRD focused too much on near-term, mission
focused objects.
In response, a provision in the America COMPETES Act, legislation
that was drafted by this committee and now public law, aims to ensure
that the NITRD supports large scale applications.
I look forward to hearing more today from our witnesses about what
other legislative changes are necessary to the NITRD program.
I yield back.
Chairman Gordon. At this time I would like to introduce our
witnesses, and we have a very distinguished group here. First,
Dr. Chris Greer is the Director of NITRD National Coordination
Office. Welcome. Dr. Daniel Reed is the Director of Scalable
and Multicore Computing at Microsoft. Dr. Craig Stewart is
Associate Dean of Research Technologies at Indiana University.
He is representing the Coalition for Academic Scientific
Computation. And finally, Dr. Don Winter is the Vice President
of Engineering and Information Technology for the Phantom
Works, Boeing Company.
Our witnesses should note we try to limit our spoken
testimony to five minutes each which then Members will have an
opportunity to question for five minutes. But this is an
important topic. We have Members at a variety of other hearings
today, and so we want to be sure we get all the information so
please feel free to again, if you need a few more minutes to
give us what you think is best, we want to hear from that.
At this point, let us start with Dr. Greer.
STATEMENT OF DR. CHRISTOPHER L. GREER, DIRECTOR, NATIONAL
COORDINATION OFFICE FOR NETWORKING AND INFORMATION TECHNOLOGY
RESEARCH AND DEVELOPMENT (NCO/NITRD)
Dr. Greer. Good morning. I am Chris Greer, and I am
Director of the National Coordination Office for Networking and
Information Technology Research and Development, the NITRD
Program. I want to thank Chairman Gordon, Ranking Member Hall,
and the Members of the Committee for the opportunity to come
here today to discuss this important issue with you. I am also
accompanied by a number of members seated behind me of the NCO
NITRD staff, and it is an honor to sit here and represent the
remarkable work of that group. And I commend this committee for
initiating the High-Performance Computing Act of 1991 and its
subsequent amendments, with its remarkably far-sighted mandate
for R&D coordination. The resulting federal program now in its
17th year has become a model for multi-agency cooperation. The
program has grown substantially in size and scope since 1991.
Today NITRD encompasses $3.5 billion in R&D across 13 member
agencies as of the President's 2009 budget.
NITRD investments further our nation's goals for national
defense and national security, health care, energy, education,
economic competitiveness, environmental sustainability, and
other national priorities.
My written testimony provides detailed responses to the
questions the Committee asked me to address. In my remarks
today, I want to highlight strategic planning. This is NITRD's
main program-wide coordination activity, and Appendix 3 of that
written testimony provides a detailed timeline for that
strategic planning and road mapping process. But in fact, even
if the PCAST, the President's Council Advisors on Science and
Technology, hadn't recommended that NITRD develop a strategic
plan, it would still be the right time in the program's history
to consider where NITRD is going and how we can better manage
that journey.
The networking and IT landscape is shifting rapidly, and
major new national challenges are emerging. The program has
recently been tasked, for example, with expanded coordination
responsibilities under the federal plan for advanced networking
R&D, and the Administration's comprehensive National Cyber
Security Initiative. The PCAST assessment of NITRD with its
provocative and important focus on rising global competition
for networking and IT leadership sharpens our thinking about
the role of strategic planning and shaping NITRD growth and
change to meet national needs. The PCAST report's 17
recommendations, seven of which go to the issue of planning,
provide a unique opportunity to make progress toward our goals.
The NITRD Subcommittee last November approved development of a
comprehensive strategic plan for NITRD and authorized my office
to add a technical staff member dedicated to support of that
activity.
The NITRD Subcommittee has agreed that the plan should be
first vision driven with the theme of complexity and multiple
dimensions. Second, focused on goals and capabilities that can
only be achieved through interagency cooperation and
coordination and the R&D capabilities and challenges required
to achieve those goals. It should also be predictive of an
effective organizational structure for the NITRD program.
As shown on the timeline in my written testimony, NITRD's
strategic planning task group is working intensively on the
plan now, including steps to solicit broad, private-sector
input to the planning process.
A request for input has just now been published in the
Federal Register and widely disseminated to academia, to
industry, and to professional organizations. A national
workshop planned for November 2008 will provide a second
opportunity for public input, and the final draft will be
posted for public comment in early 2009.
That timeline also shows that PCAST recommendations on
assessing the alignment of the NITRD research areas initiating
an NCO plan to support the overall planning process, and
preparing a fast-track education study are also being
addressed.
Other PCAST recommendations will be integrated into the
NITRD planning enterprise as the comprehensive strategic plan
takes shape and provides the larger context. Upon completion of
this strategic plan, we anticipate providing a point-by-point
response to the PCAST recommendations informed and supported by
that plan.
We agree with PCAST that leadership in networking and
information technology is essential to U.S. economic
prosperity, security, and quality of life. The federal
investments that we make in research and development in this
area are the keys to a future of promise for our nation and for
its citizens.
I look forward to working with the Congress to fulfill that
promise. Thank you.
[The prepared statement of Dr. Greer follows:]
Prepared Statement of Christopher L. Greer
Good morning. My name is Chris Greer and I am Director of the
National Coordination Office (NCO) for Networking and Information
Technology Research and Development (NITRD). With my colleague, Dr.
Jeannette Wing of the National Science Foundation (NSF), I co-chair the
NITRD Subcommittee of the National Science and Technology Council's
(NSTC) Committee on Technology. I want to thank Chairman Gordon,
Ranking Member Hall, and the Members of the Committee for the
opportunity to come before you today to discuss the Federal
Government's multi-agency NITRD effort.
The NITRD Program--now in its 17th year--represents the Federal
Government's portfolio of unclassified investments in fundamental,
long-term research and development (R&D) in advanced networking and
information technology (IT), including high-end computing, large-scale
networking, cyber security and information assurance, human-computer
interaction, information management, high-confidence software and
systems, software design, and socioeconomic, education and workforce
implications of IT. NITRD research is performed in universities,
federal research centers and laboratories, federally funded R&D
centers, private companies, and nonprofit organizations across the
country. Agencies participating in the NITRD program--including 13
member agencies and a number of other participating agencies and
offices--support vital investments in R&D and research infrastructure
to further our nation's goals for national defense and national
security, health care, energy, education, economic competitiveness,
environmental sustainability, and other national priorities. Through
the NITRD program, federal agencies work together to ensure that the
impact of their efforts is greater than the sum of their individual
investments. This is accomplished through interaction across the
government, academic, commercial, and international sectors using
cooperation, coordination, information sharing, and joint planning to
identify critical needs, avoid duplication of effort, maximize resource
sharing, and partner in investments to pursue higher-level goals.
Mandate for coordination
Seventeen years ago, when Congress passed the bipartisan High-
Performance Computing (HPC) Act of 1991 (Public Law 102-194), the Act's
mandate for interagency coordination of federal networking and IT R&D
was remarkably farsighted. The Act established a powerful, resilient
framework for federal networking and IT R&D activities. That framework
combined ambitious research goals with specific requirements for
interagency cooperation, collaboration, and partnerships with industry
and academia. The validation of the HPC Act's core vision can be found
in the success and vitality of today's NITRD Program, which has become
a model for coordination across federal agencies.
The HPC Act was amended by the Next Generation Internet Research
Act of 1998 (Public Law 105-305) and the America COMPETES Act of 2007
(Public Law 110-69). These Acts extended the scope of responsibilities
for interagency coordination to include human-centered systems;
flexible, extensible, inter-operable, and accessible network
technologies and implementations; education, training, and human
resources; and other areas. As a result, the NITRD Program now provides
for cooperation and coordination across a broad landscape, allowing it
to tackle the inherently multi-disciplinary, multi-technology, and
multi-sector challenges of today's networking and IT research horizons.
The Office of Science and Technology Policy (OSTP), with the
support of the Office of Management and Budget (OMB) and the
participating NITRD agencies, has taken a vigorous approach to
implementing the enabling NITRD legislation. The NCO Director is a
member of the OSTP technical staff group with direct access to and
active support by OSTP and OMB staff and leadership. In addition to
their financial contributions, the participating agencies provide the
time of some of their most capable experts and senior managers to
pursue NITRD goals. The success of NITRD is due in large measure to the
dedication and commitment of those who implement the program.
Program history in brief
In its first annual report to the Congress, the then-High
Performance Computing and Communications (HPCC) Program reported an
estimated 1991 multi-agency budget of $489.4 million and a proposed
1992 budget of $638.3 million. Eight federal agencies were represented
in that budget: Defense Advanced Research Projects Agency (DARPA),
Department of Energy (DOE), Environmental Protection Agency (EPA),
National Aeronautics and Space Administration (NASA), National
Institutes of Health (NIH), National Institute of Standards and
Technology (NIST), National Oceanic and Atmospheric Administration
(NOAA), and National Science Foundation (NSF). The HPCC program had
four major research areas called Program Component Areas (PCAs): High
Performance Computing Systems (HPCS); Advanced Software Technology and
Algorithms (ASTA); National Research and Education Network (NREN); and
Basic Research and Human Resources (BRHR).
Since 1991, the Federal IT R&D program has evolved continuously,
addressing the continuing, dramatic expansion in computing and
networking technologies, applications, and societal needs by adjusting
the research focus and adding new, emerging areas of interest. This
includes disaggregating investments in high-end computing
infrastructure and applications from those in high-end computing (HEC)
systems and system software research, and adding software design and
productivity, high-confidence software and systems, and societal and
workforce implications of IT.
Today, the NITRD Program, which is successor to the original HPCC
Program, encompasses $3.5 billion (2009 Budget Request) in R&D funding
and comprises 13 member agencies--the original eight agencies plus
Agency for Healthcare Research and Quality (AHRQ), National Archives
and Records Administration (NARA), Department of Energy/National
Nuclear Security Administration (DOE/NNSA), National Security Agency
(NSA), and Office of the Secretary of Defense and Department of Defense
Service research organizations (OSD and DOD Service research
organizations). About a dozen other agencies that are not formal NITRD
members also participate in the eight Program Component Areas (PCAs)
and other NITRD activities. (See Appendix 1 on page 14 for a list of
the current NITRD agencies and PCAs and a NITRD organizational chart.)
Response to the Committee Request
The invitation to testify from this House Committee included a
request to address three topic areas. Responses are provided in the
numbered sections that follow.
Request #1: Current planning and coordination overview
The NITRD Program uses five general mechanisms to pursue its
mission:
(1) Monthly meetings of the seven Federal Interagency Working
Groups (IWGs) and Coordinating Groups (CGs) chartered under the
auspices of the NSTC
(2) Workshops, most including private-sector as well as
federal participants
(3) Formal reports, including the annual NITRD Supplement to
the President's Budget and strategic planning documents
(4) Support for external studies and assessments
(5) Outreach to the federal and private sectors
I'll illustrate how these are used with examples for each
mechanism.
In each NITRD Program Component Area (PCA), agencies work together
in a CG or IWG that meets monthly to identify research needs, plan
programs, share best practices, and review progress. These regular
meetings allow groups to explore complex research and development
challenges. As an example, the High Confidence Software and Systems
(HCSS) CG is playing a leadership role in engaging researchers and
industry in assessing the national research needs in the complex life-
and safety-critical technologies called cyber-physical systems\1\
(defined here as IT embedded in and critical to the function of a
physical system; aircraft avionics are an example). This analysis is
being informed by a workshop series engaging the academic, commercial,
and government sectors. Recent workshops in this series covered medical
device software and systems, with participation by researchers,
clinicians, hospital administrators, and industry representatives;
another focused on automotive safety, engaging automobile designers,
safety experts, and engineers and academic researchers. The next in the
series, planned for Fall 2008, will focus on ``High Confidence Cyber-
Physical Transportation Systems: A look at the Commercial Aero, Auto,
and Rail Sectors, and Military Ground and Aerial Unmanned Autonomous
Vehicles (UAVs).''
---------------------------------------------------------------------------
\1\ In its 2007 assessment of the NITRD Program, the President's
Council of Advisors on Science and Technology (PCAST) termed cyber-
physical systems ``a national priority for Federal R&D.''
---------------------------------------------------------------------------
During the 20-month period from October 2006 to May 2008, the NITRD
Program planned and held a total of 27 workshops--an average of 1.5
workshops per month. Topics include composable cyber systems,
supervisory control and data acquisition (SCADA) systems for industrial
process/system control, and an upcoming event on national and
international networking research challenges. An ongoing series, the
Collaborative Expedition Workshops, covers wide-ranging topics such as
virtual work settings, evaluation of emerging technology and technology
development programs, and scalable data management.
Formal reports produced during this same 20-month period include
the 2007 and 2008 NITRD Supplement to the President's Budget and the
following strategic planning documents produced by ad hoc interagency
task groups of NITRD member agencies and others:
Federal Plan for Advanced Networking Research and Development
On January 30, 2007, OSTP Director Jack Marburger established
an Interagency Task Force on Advanced Networking and charged it
with developing a strategic vision and long-range plan for
federal networking R&D; he requested that the initial draft of
the plan be completed in three months, by May 2007, to provide
timely input for the FY 2009 budget process. Through intensive
efforts, the 40-member task force of NITRD and other agency
representatives produced a draft on schedule, including a
detailed analysis of networking research challenges that has
been extremely well received. The Task Force continued to
refine the draft over the next 12 months; the final Federal
Plan for Advanced Networking Research and Development is now
being printed and will be sent to all Members of Congress
shortly. The preprint version of the Plan is available on the
NITRD Web site at: http://www.nitrd.gov/ITFAN-preprint-
061108.pdf
Plan for Coordination of Federal R&D and Plan for the Leap-
Ahead Program of Research and Development
In February 2008, OSTP called for an Interagency Task Force
from NITRD agencies and others to develop two research-related
planning documents on a fast-track basis under the
Comprehensive National Cybersecurity Initiative (CNCI),
established by National Security Presidential Directive 54/
Homeland Security Presidential Directive 23 in January 2008. To
expedite quick turnaround on this tasking, the 21 task force
members divided into two groups. One developed the plan for
overall coordination of the federal cyber R&D portfolio; the
other crafted the ``Leap-Ahead'' plan for accelerating high-
risk, high-return research to help maintain our technological
edge in cyberspace. These plans now provide the basis for the
recent launch of the CNCI R&D planning activities.
Under the CNCI plans, the Cyber Security and Information
Assurance Interagency Working Group (CSIA IWG) chartered by the
NSTC in 2006--augmented by a new Senior Steering Group--is
tasked with two new assignments: leading the CNCI R&D
coordination activity including improving coordination between
the unclassified and classified Federal R&D sectors, and
coordinating the ``Leap-Ahead'' initiative. The CSIA IWG's 2006
Federal Plan for Cyber Security and Information Assurance
Research and Development provides a detailed technical baseline
for setting federal cyber R&D priorities under CNCI.
The NITRD Program supports external studies and reviews to expand
its perspectives and take advantage of expertise from a diversity of
sources. A study by the National Academies is currently underway to
develop a better understanding of the potential scientific and
technological impact of high-end capability computing in science and
engineering. Public release of the final report is expected in
September 2008. The Program recently provided briefings and written
inputs to the Networking and Information Technology Subcommittee of the
President's Council of Advisors on Science and Technology (PCAST) for
use in its assessment of the NITRD Program. Looking ahead, the Program
developed a statement of work for the first of the fast-track studies
on NIT post-secondary education called for by the PCAST assessment of
NITRD.
The NITRD Program uses a variety of mechanisms to reach out to
researchers, private-sector developers, resource providers, and end-
users. Examples include two groups under the Large Scale Networking CG:
the Joint Engineering Team (JET) and Middleware and Grid Infrastructure
Coordination (MAGIC) group, which have academic and industry members;
the Federal Agency Administration of Science and Technology Education
and Research (FASTER) Community of Practice (CoP), which seeks
exchanges of information with the private sector and new technologies
to streamline the management of federal research; and the multi-sector
NITRD research workshops held in all the PCAs.
A number of efforts are underway to improve the effectiveness of
NITRD planning and coordination. These include revamping the NITRD web
site (both public and federal-only resources), providing improved web-
based services to support remote participation and digital content
sharing, and outreach visits by NCO technical staff to academic and
commercial organizations as a required component of regular conference
travel.
The high sustained level of collaborative engagement reflected in
the diverse NITRD activities of the last two years is, in my judgment,
a key measure of the effectiveness of the NITRD coordination model--it
remains resilient amid the Program's increasing activities and
expanding responsibilities. Another measure is the productive synergy
gained through joint funding, partnerships with private-sector
entities, and sometimes a combination of the two. For example:
Collaboration
Benchmarks for Federal HEC systems: The HEC agencies are collaborating
to develop an interagency suite of HEC benchmarks that can accurately
represent the demands of federal advanced computing applications.
IPv6 debugging: DOD, DOE/SC, and NASA are collaborating, in cooperation
with the university networking consortium Internet2, in a project that
is conducting end-to-end debugging, performance measurement, and
toolset enhancement of Internet Protocol version 6 (IPv6) over DOD's
Defense Research and Education Network (DREN), DOE/SC's Energy Sciences
network (ESnet), and Internet2Net.
Environmental databases and data distribution: Through the Earth System
Modeling Framework activity and related efforts, NITRD agencies (DOD,
EPA, NASA, NOAA, NSF) continue their long-range cooperative work to
expand the inter-operability and usability of diverse models and large-
scale data sets for weather, climate, and environmental research.
Joint funding/Partnerships
High-Productivity Computing Systems (HPCS) Phase III: This DARPA
effort, supported also by DOE/SC and NSA and with collaborative
participation by other HEC agencies, involves design, fabrication,
integration, and demonstration of full-scale prototypes by 2010 for a
new generation of petascale, economically viable computing systems.
HEC-University Research Activity (HEC-URA): In 2004, HEC R&D agencies
(DARPA, DOE/NNSA, DOE/SC, NASA, NSA, and NSF) initiated this program of
high-risk R&D in technically challenging areas including HEC software
tools and compilers; file systems, I/O, and storage design for high
throughput; and new parallel programming models for thousands of
processors. DARPA, DOE/SC, and NSF have contributed funding, and they
and other HEC agencies participate in reviews and HEC-URA workshops.
DETERlab: DHS and NSF, with university and industry partners, are
supporting the cyber-DEfense Technology Experimental Research
laboratory testbed, a general-purpose experimental infrastructure that
enables research and development on next-generation cyber security
technologies.
Open Science Grid (OSG): NSF and DOE/SC are jointly supporting this
growing consortium of about 100 researchers and software, service, and
resource providers from universities, national laboratories, and
computing centers across the U.S. OSG brings together distributed
computing and storage resources from campuses and research communities
in a common, shared grid infrastructure over research networks via a
common set of middleware.
Trustworthy Cyber Infrastructure for the Power Grid (TCIP): In this
effort co-funded by NSF, DOE (OE), and DHS, researchers from the
University of Illinois at Urbana-Champaign, Dartmouth College, Cornell
University, and Washington State University are seeking to better
secure operations of the Nation's power grid by improving the
engineering of its underlying IT infrastructure, making it more secure,
reliable, and safe.
Cluster Exploratory (CluE) program: NSF has formed a partnership with
Google and IBM that will enable academic researchers to explore data-
intensive computing applications in science and engineering using a
1,600-processor server farm set up and supported by the two companies.
Committee Request #2: PCAST assessment of the NITRD Program
Periodic assessments of the multi-agency networking and IT R&D
program by a Presidential advisory committee are mandated by the HPC
Act, as amended by the Next Generation Internet Research Act of 1998
and most recently by the America COMPETES Act of 2007. Executive Order
13385, signed September 29, 2005, assigned the assessment
responsibility to PCAST, which in 2006 established a Networking and
Information Technology (NIT) Subcommittee to lead the process. The
results of PCAST's assessment are presented in the August 2007 report
Leadership Under Challenge: Information Technology R&D in a Competitive
World.
Over all, the PCAST concluded that while the NITRD program, with
NCO support, has in the past ``been effective at meeting agency and
national needs,'' for the future ``changes are needed in order for the
United States to ensure its continued leadership.'' This conclusion
recognizes the advent of an era of global NIT competitiveness in which
``other countries and regions have also recognized the value of NIT
leadership and are mounting challenges.'' The changes recommended by
PCAST are in the areas of education and workforce development,
portfolio balance, new emphasis areas, and strategic planning. The
PCAST conclusions and recommendations sharpen our focus on the central
role of strategic planning in shaping NITRD growth and change; and even
in the most technically difficult R&D areas such as complex software,
the PCAST recommendations provide an opportunity to make progress
toward our goals.
The PCAST makes 17 recommendations in its report. (The
recommendations are listed numerically, in sequence by chapter, in
Appendix 2. Recommendations are noted parenthetically by number in this
testimony.) These recommendations can be categorized as follows:
(1) Seven focus on improved planning processes (#9, 11-13, 16,
18, 20)
(2) Four address issues of portfolio balance and emphasis
areas (#2a, 6, 8, 14)
(3) Two suggest studies or consultations (#1, 10)
(4) Two focus on assessment (#17, 19)
(5) Two are addressed to the Director of OSTP (#7, 15)
(6) Three call for efforts to ease the visa process for
international students, graduates, and visiting experts (#2b,
2c, 2d)
The final two categories fall outside the purview of NITRD and this
testimony, and will not be addressed further. I would like to address
the first four categories with a few comments and observations on each.
PCAST Category 1: Planning recommendations
The PCAST assessment comes at a developmental turning point for
NITRD. In light of the maturation and increase in responsibilities I
have described, it is clearly the right time in NITRD history to
consider where we are going and how we can better manage the journey.
For this reason, and in light of the PCAST assessment, the NITRD
Subcommittee has initiated the development of a comprehensive strategic
plan. The key features of this plan are that it is:
Vision-driven with a theme of complexity in multiple
dimensions
Focused on goals and capabilities that can only be
achieved through interagency cooperation and coordination, and
the R&D capabilities and challenges required to achieve those
goals
Predictive of an effective organizational structure
for the NITRD Program
With the development of a comprehensive strategic plan, we
anticipate a point-by-point response to the PCAST recommendations
informed and supported by the plan.
Process for developing NITRD Strategic Plan
At its November 2007 meeting, the NITRD Subcommittee approved an
initiative to prepare a new Strategic Plan for NITRD as the critical
initial task for entering a new phase of development. A detailed
timeline for the strategic planning process, with milestones, is
provided in Appendix 3 (note that the timeline also lists the PCAST
recommendations relevant to the various steps in the process). This
timeline covers the period FY 2008-09 and has five major features:
(1) The plan development process has three subphases--initial
content development March through September 2008; text drafting
and revision September 2008 through March 2009; and concurrence
review with a target for release in June 2009.
(2) The process provides multiple opportunities and mechanisms
for public input including a Request for Input (RFI) for
initial comments, a workshop to engage all sectors, and public
comments on a full draft plan.
(3) The PCAST recommendations are fully integrated into and
help guide the strategic planning process.
(4) The development of PCA strategic plans and roadmaps
overlaps with and is informed by the culmination of the NITRD
strategic planning process.
(5) The strategic planning process is viewed as ongoing with
regular opportunities in the future for evolving and revising
the plan as goals are achieved and the networking and IT
landscape changes.
Agency representatives kicked off the strategic planning process
with a two-day off-site meeting in March 2008. First principles agreed
upon at that meeting were that the NITRD Strategic Plan should align
with the strategic plans of the member agencies, and that the Plan
should focus on long-term capabilities that require the research
contributions of multiple agencies to achieve. An 18-member strategic
planning team of agency representatives is now meeting weekly and is
currently focused on the task of initial content development. A Request
for Input (RFI) appeared in the Federal Register July 24 and
notification has been sent to stakeholder organizations across the
country as well as to the NCO's outreach list of approximately 1,700
contacts. The two-page RFI (see Appendix 3) asks all interested
parties--individuals, groups, organizations, and representatives of
companies and industries--to provide a two-page statement envisioning
the future of networking and IT and the future role of NITRD.
In developing its strategic plan, NITRD is also coordinating
closely with the NSTC Committee on Science's Interagency Working Group
on Digital Data (IWGDD). The IWGDD is charged with developing and
providing for the implementation of a plan to cultivate a framework for
reliable preservation and effective access to digital scientific data.
Along with Cita Furlani of NIST and Charles Romine of OSTP, I serve as
co-chair of the IWGDD.
PCAST Category 2: Recommendations on portfolio balance and emphasis
areas
This category of PCAST recommendations recognizes and supports the
current NITRD portfolio while suggesting increases in:
(1) larger-scale, longer-term, multi-disciplinary, and high-
risk/high-payoff research; and
(2) support for NIT systems connected with the physical world,
software, digital data, and networking, while continuing
support for high-end computing, cyber security and information
assurance, human-computer interaction, and NIT and the social
sciences.
As PCAST recognizes, the NITRD Program fields a number of efforts
in this first item today, including R&D in petascale architectures,
software, and applications; all-optical network technologies; quantum
information technologies; and next-generation wireless and sensor
capabilities. At the same time, a key goal of NITRD's current strategic
planning activity is to enable us to identify new opportunities for
long-term, high-risk research investments. The plan's specific emphasis
on goals and capabilities that can only be achieved by agencies working
together is intended to enable agencies to share funding for larger and
longer-term projects and to share the risk in projects whose payoffs
are broad enough to interest multiple agencies. Furthermore, the
Program's ability to move nimbly to seize such new opportunities is
contingent in part on the alignment of the PCAs in which agencies
report their NITRD research dollars. For that reason, one focus of our
strategic planning activities is an unfettered examination of the PCAs
to assess whether, and what type of, realignment of NITRD research
areas might be desirable to promote new strategic directions. (This
kind of Subcommittee assessment is also called for by the PCAST in a
separate recommendation.)
High payoffs can also come from good ideas that are not necessarily
high-risk. Two such examples are the opening up of computing cycles on
federal leadership-class systems to the broader national research
community and the investment by NSF in Track 2 HEC clusters. The NSF
investment resulted in a dramatic increase in computational resources
available over the Teragrid. The open solicitations for leading-edge
computational research proposals by DOE/SC (under the Innovative and
Novel Computational Impact on Theory and Experiment [INCITE] program)
and NASA (under the former National Leadership Computing System [NCLS]
program) have greatly broadened access for the national research
community to the world's most powerful supercomputers. The 2008 INCITE
competition resulted in awards of computing cycles on leadership-class
federal systems to eight major U.S. corporations, 17 universities, and
20 smaller federal agencies and labs as well as international research
institutions--for a total of more than a quarter of a billion compute
hours.
The topic areas listed in the second item above (focused on cyber-
physical systems) are emerging as crucial in the discussions of the
NITRD strategic planning group. We concur with PCAST in its assessment
of the importance of these topics and expect them to be central in the
final strategic plan.
Although the PCAST report states that ``over all, technology
transfer has worked well in networking and IT,'' the NITRD Program has
several new opportunities to address the report's recommendation that
NITRD do more to exploit existing tech transfer mechanisms. Already
existing NITRD mechanisms that bring researchers and their results
together with private-sector developers and end-users include: the
above mentioned JET and MAGIC groups; the Federal Agency Administration
of Science and Technology Education and Research (FASTER) community of
practice group, which seeks exchanges of information with the private
sector and new technologies to streamline the management of federal
research; and the multi-sector NITRD research workshops held in all the
PCAs.
The new opportunities are presented by the two CNCI plans and the
advanced networking plan. Each of these plans places substantial
emphasis on developing new models for expanding substantive
interactions with the private sector, such as cooperation on testbeds
and increased meetings with industry organizations, and on expediting
the movement of research results into prototyping and commercial
implementation. The increasing pace of technological change is
recognized in the NITRD community as a challenge in advancing research
innovations, so there is eagerness now to explore ways to improve
NITRD's outreach to private developers and industry.
The new CNCI activities also bear on the PCAST recommendation to
increase the emphasis on long-term research and infrastructure in cyber
security and information assurance. The NITRD Subcommittee has approved
the addition of one FTE to the NCO staff to support the expanded
responsibilities of the CSIA IWG and its new Senior Steering Group
(SSG) for coordinating cyber R&D and the Leap-Ahead research
initiative. Infrastructure for cyber security R&D is called for by both
the CNCI planning documents and the CSIA IWG Federal Plan.
PCAST Category 3: Recommendations for consultations and studies
The dynamic and global networking and IT landscape will require a
partnership across the government, academic, and commercial sectors if
we are to maintain our nation's leadership role. This will require the
Federal Government to act as both leader to and partner with the other
sectors. The NITRD agencies can lead by making effective R&D
investments, including those in larger, longer-term, multi-
disciplinary, and high-risk/high-payoff efforts, and by setting
examples, demonstrating feasibility, and developing initial
implementation capabilities through their own NIT activities, such as
achieving IPv6 capability. The NITRD agencies can be partners by being
transparent and interactive in their R&D planning activities,
exchanging information about emerging innovations and understanding the
needs, opportunities, and capabilities in the other sectors.
This dual leadership/partnership role requires ongoing mechanisms
for dialogue and interaction between the NITRD program and other
sectors. As I mentioned earlier, the JET and MAGIC teams include
academic and commercial-sector participation. This model could
profitably be extended into other PCAs and focus areas. The NITRD
workshops are designed to draw participation across sectors and to
bring together groups with complementary interests and capabilities
that do not have a history of interaction. This mechanism will continue
to see extensive use. The PCAST assessment and its influence on NITRD
activities demonstrates the value of high-level external review of the
Program as an additional means for input. The America COMPETES Act
calls for an ongoing, external review process.
The partnership role also includes making good use of the expertise
and perspectives available in the other sectors. External studies
commissioned by NITRD are one means for achieving this. For example,
the PCAST assessment identifies as a priority area ensuring an adequate
supply of well-educated NIT professionals, a strategic goal that we
share. To inform the development of our strategic plan, the NITRD
agencies have launched an initial fast-track study of networking and IT
education. A Statement of Work developed by a multi-agency task group
was approved at the March off-site meeting of the NITRD Subcommittee.
We are also in the process of assessing the current NITRD educational
activities including graduate fellowships to compare these against
needs and against priorities of our strategic plan. Our initial plan
includes a full-day workshop to discuss current programs across the
federal agencies. Thus, the strategic planning process itself is an
example of the use of multiple consultation and input mechanisms to
inform planning.
Additional examples of external inputs are in the areas of software
development and advanced networking. The recent National Academies
study Software for Dependable Systems: Sufficient Evidence? has been
complemented by the ongoing workshop series supported by the HCSS group
that has drawn input from academia, industry, user groups, and
government on certifiably dependable software systems for critical
applications. The Federal Plan for Advanced Networking Research and
Development was informed by a series of eight workshops, RFIs, working
groups, and external reports.
PCAST Category 4: Recommendations on assessment
The PCAST assessment included recommendations for periodic
assessment of the NITRD PCA structure and the development of metrics
and indicators to assess progress. As I stated earlier, an explicit
goal in the strategic planning process is to evaluate the current PCA
structure against our new strategic plan and to make changes as
appropriate. We envision the strategic planning process and any
associated PCA realignments as an ongoing process, to be revisited on a
regular basis as the networking and IT landscape evolves and as
strategic goals are achieved.
There are currently two types of metrics or indicators against
which we intend to assess progress. Stage One indicators include
successful completion of the Strategic Plan and the PCA strategic plans
and roadmaps--including measures of progress--called for in the PCAST
report. The timeline in Appendix 3 provides a series of specific
milestones and events, which are examples of Stage One indicators.
Stage Two indicators--measures of how well the Program is carrying out
its Strategic Plan, how effectively the PCAs are pursuing their
strategic plans and roadmaps, and the impact of these efforts--are
being developed as part of the strategic planning process. These Stage
Two indicators will be an important part of our implementation plan.
Committee Request #3: Role and functions of the National Coordination
Office for NITRD (NCO/NITRD)
The NCO/NITRD is identified in the NSTC Committee on Technology
charter for the NITRD Subcommittee. The Office provides technical,
planning, budgetary, and logistical support for all the activities of
the NITRD Program, under the operative framework of relevant laws,
charters, and Executive Branch directives. The Office also serves as
the central point of contact for inquiries and requests for information
about the Program and maintains the Program's Web site and documents,
including current and archival documentation of NITRD subcommittee,
IWG, and CG activities. The Director and Associate Director are federal
employees and serve as senior management. The staff of 13 contractors
and one federal employee on detail includes a contract manager and an
office operations manager; five Technical Coordinators who support 11
IWGs, CGs, and technical groups; one writer/editor; three
administrative support staff; a web master and an IT systems manager;
and a temporary full-time coordinator for the NITRD strategic planning
process. The five Technical Coordinators are Ph.D.-level positions that
provide expert knowledge of the R&D challenges in the NITRD fields.
Regular NCO activities include logistical preparations and staff
support for all meetings of NITRD entities, including those of the
Presidential advisory group on IT, and most NITRD-affiliated workshops;
drafting, editing, and publishing support for publications (annual
budget supplement, R&D plans, workshop reports, studies, and reviews)
of the NITRD Program and those of the Presidential advisory group; and
preparation of special budgetary and technical documents requested by
the NITRD Subcommittee. The NCO Director maintains close communications
with OSTP, OMB, the NITRD agencies, and this Committee, and represents
the Program in presentations to organizations nationally.
The PCAST assessment includes three recommendations that explicitly
reference the NCO. Two focus on NCO support for the Subcommittee in
commissioning studies on networking and IT education and in developing
metrics and progress indicators for assessment. These support efforts
are underway, as described above.
The third recommendation is that NCO, with Subcommittee guidance,
should develop and implement a plan for supporting the NITRD Program in
developing strategic plans and roadmaps. Such a plan has been developed
for the initial stages of this new NITRD activity and is being
implemented. Under this initial plan, the NCO has committed significant
resources to the process, including the hiring of a temporary
coordinator for strategic planning. The Office has committed
significant technical writing time in preparing text and has charged
the Technical Coordinators with serving as liaisons between the
Strategic Planning Group and the IWGs and CGs. The Office is supporting
the weekly meetings of the Strategic Planning Group and providing
logistical support for its outreach activities. Thus, the NCO is fully
committed to supporting a successful NITRD strategic planning and
roadmapping process.
In conclusion
The enabling NITRD legislation and its vigorous implementation by
OSTP, OMB, and the NITRD agencies have created a robust, responsive,
and resilient framework for effective cooperation and coordination in
federal networking and IT R&D planning and execution. The NITRD Program
has matured and now encompasses a spectrum of NIT areas that allow it
to take on the complex, multi-disciplinary, multi-sector challenges
characteristic of today's networking and IT landscape.
With this maturation comes the opportunity and responsibility for
comprehensive strategic planning to ensure best use of this important
resource for coordination. The NITRD Program is now deep into the
process of a vigorous strategic planning and roadmapping effort. We are
confident that this process and its attendant elements will fully
address the valuable recommendations contained in the PCAST assessment.
A measure of the strength of the NITRD Program and the supporting
National Coordination Office is the ability to simultaneously support a
vigorous strategic planning process, the development of coordination
and leap-ahead R&D activities under the Comprehensive National
Cybersecurity Initiative, manage two external studies, facilitate a
robust workshop series, and conduct the regular planning, coordinating,
and reporting activities of the 11 IWGs, CGs, and teams. This is only
accomplished because of the competence, dedication, and commitment of
all of the members of the NITRD/NCO community.
As the PCAST concludes, leadership in networking and information
technology is essential to U.S. economic prosperity, security, and
quality of life. The federal investments we make in research and
development in this area are the keys to a future of promise for our
nation and its citizens. I look forward to working with Congress to
fulfill that promise.
Thank you.
Biography for Christopher L. Greer
Dr. Chris Greer is Director of the National Coordination Office
(NCO) for the Networking and Information Technology Research and
Development (NITRD) program. The NCO/NITRD mission is to formulate and
promote federal information technology research and development to meet
national goals. The NCO reports to the Office of Science and Technology
Policy within the Executive Office of the President. Dr. Greer is on
assignment to the NCO from his position as Senior Advisor for Digital
Data in the NSF Office of Cyberinfrastructure. He recently served as
Executive Secretary for the Long-lived Digital Data Collections
Activities of the National Science Board and is currently Co-Chair of
the Interagency Working Group on Digital Data of the National Science
and Technology Council's Committee on Science. He is also a member of
the Advisory Committee for the National Archives and Records
Administration's Electronic Records Archive and a member of the Digital
Library Council of the Federal Depository Library Program.
Dr. Greer received his Ph.D. degree in biochemistry from the
University of California, Berkeley and did his postdoctoral work at
CalTech. He was a member of the faculty at the University of California
at Irvine in the Department of Biological Chemistry for approximately
18 years where his research on gene expression pathways was supported
by grants from the NSF, NIH and the American Heart Association. During
that time, he was founding Executive Officer of the RNA Society, an
international professional organization.
Chairman Gordon. Right on the money, Dr. Greer. Dr. Reed,
you are up.
STATEMENT OF DR. DANIEL A. REED, DIRECTOR, SCALABLE AND
MULTICORE COMPUTING, MICROSOFT CORPORATION
Dr. Reed. Good morning, Mr. Chairman and Members of the
Committee. I am Dan Reed. I am Chair of the Board of Directors
of the Computing Research Association and Co-Chair of the PCAST
Subcommittee that produced the 2007 NITRD Program Assessment.
Today I would like to make five points regarding the NITRD
Program followed by a set of specific recommendations for the
future. Information technology, as Dr. Greer noted, is driven
by basic research investments that has transformed our society
and our economy. Imagine a world without personal computers,
without mobile devices or the Internet, without predictive
computational models or deep--the future can be even more
amazing if we sustain our IT research----
Historically, the diversity of the NITRD agencies has been
a major strength of U.S. IT research, fostering multiple
approaches to complex problems. The Internet began as a DARPA
project, grew with NSF support, and blossomed with commercial
funding. The human genome project was a triumph of biomedicine
and IT based and building on funding from NIH, from DARPA, from
NSF, the Department of Energy, and it birthed personalized
medicine.
This brings me to the issue of balancing risk and
protection--today I believe the NITRD ecosystem's health is
threatened due to an over-dependence on a single-funding source
and inadequate research funding overall. DARPA's retreat from
fundamental computing research at U.S. universities has
unbalanced the NITRD ecosystem. NSF now provides 86 percent of
all academic IT research funding, and fierce competition has
driven researchers to focus excessively on short-term, well-
risked research projects. Like a stock portfolio, our long-term
success depends on balance, planning, and regular reassessment.
This leads to my third point, NITRD coordination and
planning. In general, I believe the NITRD Programs effectively
foster informal communication and coordination among the
agencies, and I commend the National Coordination Office for
its role in this aspect. However, the focus on individual
agency agendas has made the NITRD Program less effective in
managing coordinated projects, particularly multi-disciplinary
ones of rising importance.
This leads to my fourth point about research opportunities
and foci. In 2007, PCAST revisited the priority areas
identified by PITAC in 1999. Concluding that they remained
deeply relevant, including I should add, as a personal
anecdote, high performance computing, something which I have
been involved in for many years. IT systems that interact with
the physical world, however, a special case is the more general
issue of software systems emerged as a new top priority. These
cyber-physical systems embed computing, sensors, and actuators
and objects that span scales from our critical national
infrastructure to implanted biomedical devices. Their creation
requires workers with new and ever-more multi-disciplinary
skills.
That leads me to the issue of sustaining our IT workforce.
Today information technology has a serious image problem. It
affects our workforce quantity, its diversity, and its--many
groups are working very hard to address stereotypes and create
new, multi-disciplinary curricula but much work remains in this
area. I believe we must also do more to retain the best and
brightest international students who obtain graduate degrees
here, many of whom are supported by federal research grants and
contracts. Simply put, our international competitiveness
depends on the availability of a qualified and diverse
workforce.
This leads to my recommendations for the future. To ensure
the health of the U.S. IT ecosystem, we should fully fund the
America COMPETES Act. This will fuel the IT innovation engine,
the fundamental research by U.S. universities and national
laboratories, and further broaden STEM-based education. And I
commend you, Mr. Chairman, and your colleagues for your
continuing support of America COMPETES.
Second, I believe we must rebalance participation in the
NITRD Program so that responsibility for fundamental research
is not born excessively by a single agency. As Dr. Greer noted,
I believe we must create and regularly update a strategic R&D
plan and a set of associated metrics that define interagency
accountabilities with a mix appropriately of project scales
and--risks.
Finally, I believe we must regularly review our research
investment against that strategic plan. I also believe the
NITRD Program is best served by a stand-alone and active PITAC
that is composed of computing experts drawn from academia and
industry. I say that as someone who served on both PITAC and
PCAST. Eight years between overall NITRD reviews is far too
long in the information technology industry. By analogy, eight
years in dog years is multiple lifetimes in the computing
industry. We need to be more proactive in examining our
strategy.
Mr. Chairman, thank you and this committee for your
continued interest and support in the future of the NITRD
Program and its importance to U.S. competitiveness and national
security. At the appropriate time, I would be delighted to
answer any questions.
[The prepared statement of Dr. Reed follows:]
Prepared Statement of Daniel A. Reed
Good afternoon, Mr. Chairman and Members of the Committee. Thank
you for granting me this opportunity to comment on the federal
Networking and Information Technology Research and Development (NITRD)
program. I am Daniel Reed, Chair of the Board of Directors for the
Computing Research Association (CRA). I am also a researcher in high-
performance computing; a member of the President's Council of Advisors
on Science and Technology (PCAST); the former Head of the Department of
Computer Science at the University of Illinois at Urbana Champaign; and
currently Director of Scalable and Multicore Computing Strategy at
Microsoft.
During our lifetime, information technology has transformed our
society, our economy and our personal lives. Imagine a world without
consumer electronics, personal computers, the Internet or predictive
computational models. As Tennyson so eloquently expressed, we have ``.
. . dipped into the future, far as human eye could see; saw the vision
of the world, and all the wonder that would be.'' Despite our current
wonder, the future of computing--the world that can be--is even more
amazing, for we are poised on the brink of even greater revolutions:
deep understanding of biological and physical processes, personalized
medicine and assistive living technology, autonomous vehicles that
navigate in traffic and severe weather, strategic and tactical military
and intelligence systems with true information superiority, information
assistants that enhance our intellectual activities, distributed
sensors and actuators that protect our environment, intelligent systems
for advanced energy management, and a host of other innovations.
Making such visions a reality is the essence of information
technology research and the core of the NITRD program. It is also why
sustained and appropriate investments in information technology
research and development are critical to our nation's future.
In response to your questions, I would like to make eight points
today regarding the status and future of the NITRD program, beginning
with a synopsis of the recent report of the President's Council of
Advisors on Science and Technology (PCAST) assessment of the
Networking, Information Technology Research and Development (NITRD)
program.
1. PCAST: Information Technology Assessment
In 2007, I was privileged to co-chair PCAST's assessment of the
NITRD program. The resulting report, Leadership Under Challenge:
Information Technology R&D in a Competitive World,\1\ was the first
overall assessment of the NITRD program since that conducted in 1999 by
the President's Information Technology Advisory Committee (PITAC). The
2007 PCAST report emphasized the following points:
---------------------------------------------------------------------------
\1\ Leadership Under Challenge: Information Technology R&D in a
Competitive World, President's Council of Advisors on Science and
Technology (PCAST), August 2007, http://www.ostp.gov/pdf/
nitrd-review.pdf
NIT and global competitiveness. Today, the United
States is the global leader in networking and information
technology (NIT) and that leadership is essential to U.S.
economic prosperity, security, and quality of life. However,
other countries and regions have also recognized the value of
---------------------------------------------------------------------------
NIT leadership and are mounting challenges.
NITRD ecosystem. The NITRD program is a key mechanism
through which the Federal Government contributes to NIT
research and development leadership, and the NITRD program has
by and large been effective at meeting agency and national
needs.
Research horizons and risks. The federal NIT research
and development portfolio is currently imbalanced in favor of
low-risk projects; too many are small scale and short-term
efforts. The number of large-scale, multi-disciplinary
activities with long time horizons is limited and visionary
projects are few.
Workforce availability and skills. The number of
people completing NIT education programs and the usefulness of
that education fall short of current and projected needs.
Current curricula must be re-evaluated, graduate fellowships
increased and visa processes simplified to address these
challenges.
Research priority areas. The top priorities for new
funding are NIT systems connected to the physical world,
software, networking and digital data, with continuing emphasis
on high-end computing, cyber security and information
assurance, human-computer interaction and NIT and the social
sciences.
Strategic plans and roadmaps. We must develop,
maintain, and implement a strategic plan for the NITRD program,
along with public R&D plans or roadmaps and progress metrics
for key technical areas that require long-term interagency
coordination and engagement.
Interagency coordination. The current nature and
scale of NITRD program coordination processes are inadequate to
meet anticipated national needs and to maintain U.S. leadership
in an era of global NIT competitiveness.
With this backdrop, the remainder of my testimony expands and
explains the rationale for these PCAST findings along with personal
observations on possible actions. However, the opinions expressed
herein are my own, not necessarily those of PCAST or the Office of
Science and Technology Policy (OSTP). I would also like to acknowledge
the contributions of Peter Harsha, from the Computing Research
Association (CRA), to these remarks.
2. The Importance of Information Technology
The importance of information technology (IT) in enabling
innovation and powering the new economy is well documented. Advances in
computing and communications have led to significant improvements in
product design, development and distribution for American industry,
provided instant communications for people worldwide, and enabled new
scientific disciplines like bio-informatics and nanotechnology that
show great promise in improving a whole range of health, security, and
communications technologies. Several studies have suggested information
technology has been responsible for 25 percent of more of U.S. economic
growth in recent years, despite being a much smaller fraction of the
gross domestic product (GDP).\2\ Moreover, information technology
leadership has proven essential to the Nation's security, from our
national infrastructure and signals intelligence to our military.
---------------------------------------------------------------------------
\2\ Dale W. Jorgenson and Charles Wessner, editors. 2007. Enhancing
Productivity Growth in the Information Age: Measuring and Sustaining
the New Economy. Washington, D.C.: National Academies Press. Also see
Dale W. Jorgenson, Mun S. Ho, and Kevin J. Stiroh. 2005. Productivity
Volume 3: Information Technology and the American Growth Resurgence.
Cambridge, Mass.: MIT Press.
---------------------------------------------------------------------------
Information technology has also changed the conduct of research.
Innovations in computing and networking technologies are enabling
discovery across every scientific and engineering discipline--from
mapping the human brain to modeling climatic change and enhancing
energy production. Faced with problems that are ever more complex and
interdisciplinary in nature, researchers are using IT to collaborate
across the globe, visualize large and complex data sets, and collect
and manage massive amounts of real-time sensor-derived data.
But equally important to the role IT plays in enabling innovations
in industry and in the other scientific and engineering disciplines is
the role of the research and development (R&D) ecosystem in enabling IT
innovations. The 1995 National Research Council (NRC) report, Evolving
the High Performance Computing and Communications Initiative to Support
the Nation's Information Infrastructure,\3\ included a compelling
graphic illustrating this spectacular return. The graphic was updated
in 2002 and is reproduced in Figure 1.
---------------------------------------------------------------------------
\3\ U.S. National Research Council. Evolving the High Performance
Computing and Communications Initiative to Support the Nation's
Information Infrastructure. National Academies Press, Washington, D.C.
1995.
---------------------------------------------------------------------------
The graphic in Figure 1 shows the development of technologies from
their origins in industrial and federally-supported research, to the
introduction of the first commercial products, through the creation of
billion-dollar industries and markets. The original 1995 NRC report
identified nine of these multi-billion dollar IT industries (the
categories on the left side of the graphic). Seven years later, the
number of examples had grown to 19 multi-billion dollar industries that
are transforming our lives and driving our economy.
The graphic also illustrates the complex interplay among industrial
R&D efforts and the interdependent ecosystem of NITRD agencies that
supports academic research. Each federal agency plays a distinct, but
important role in the current and future success of the U.S.
information technology ecosystem.
3. The NITRD Ecosystem: Fostering Innovation via Diversity
The NITRD program is a collaborative confederation of thirteen
federal agencies, each with differing missions that depend--to varying
degrees--on advances in information technology. This ecosystem of
agencies is complex and interdependent, with some small and others
large, some supporting outcome-directed research and others supporting
innovation-driven research, some supporting small projects and others
funding large initiatives, some focused on federal research
laboratories and others engaging academia.
Historically, this NITRD diversity has been a major strength of the
U.S. approach to information technology research, as it has fostered
diverse approaches to complex computing problems, with differing
research horizons and communities. Together, a strong IT industry,
powerful commercialization system, and high-quality education and
research institutions have been critical to America's leadership in IT.
The aforementioned 1995 report by the National Research Council
emphasized the ``extraordinarily productive interplay of federally
funded university research, federally and privately funded industrial
research, and entrepreneurial companies founded and staffed by people
who moved back and forth between universities and industry.''
To further illustrate this point, consider some specific,
compelling examples of agency leadership, cross-agency collaboration
and industrial engagement. The Defense Advanced Research Projects
Agency (DARPA) has historically supported large-scale projects with
revolutionary intent--high-speed networks for resilient communication,
artificial intelligence and autonomous navigation, massively parallel
supercomputing for detailed modeling, real-time and embedded systems
for situational awareness--to ensure the technological superiority of
U.S. military forces. Today's Internet began in the 1960s as an
ambitious DARPA (then ARPA) research project in resilient, packet-based
communications for national defense.
Reflecting its long-term focus, DARPA supported the Arpanet for
well over a decade. This later enabled the National Science Foundation
(NSF) to build on a rich research and technology base to create a high-
speed national network connecting supercomputing centers and their NSF-
funded students and faculty researchers. From this fertile ground, the
Mosaic web browser was born at the University of Illinois, spawning the
commercial web revolution and today's Internet via commercial
investments.
The Department of Energy's Office of Science (DOE SC) and its
National Nuclear Security Administration (DOE NNSA) have long supported
algorithms and software research, network and distributed systems
studies and advanced computer architecture designs in both DOE
laboratories and academia. DOE SC's Scientific Discovery through
Advanced Computing (SciDAC) program supports multi-disciplinary teams
to develop the enabling technologies for next-generation computing
systems and their application to models of climate change, efficient
energy sources and biological processes. In turn, the DOE NNSA has
advanced computer systems, software and algorithms in support of
nuclear stockpile stewardship and certification.
The Human Genome Project, funded by the National Institutes of
Health (NIH), was enabled by high-throughput sequencing systems, based
on advanced semiconductor technology and efficient algorithms for DNA
subsequence reassembly executing atop high-performance computing
systems. Simply put, the Human Genome Project was a collaborative
triumph of biomedicine and information technology; the commercial
semiconductor designs and computer architecture and academic algorithms
that enabled this breakthrough were previously funded by DARPA, NSF and
DOE. The tantalizing promise of low-cost, personalized medicine, with
treatments and drugs tailored to individual needs, will be realized
only via continued advances in computing technology, themselves derived
from information technology research.
As all these examples illustrate, the success of the NITRD program
has accrued from the health, diversity and vigorous interactions among
its component agencies, universities and industrial partnerships.
Historically, DARPA funded large-scale, high-risk projects involving
academic and industry teams. In turn, DOE supported national laboratory
and academic researchers around large-scale scientific instruments, and
NSF supported innovation-driven research, predominantly by individual
faculty members and their students, with a mix of larger projects and
centers. NIH has partnered on selected NITRD programs and NASA, NIST
and the other NITRD agencies have supported mission-specific research
and development programs.\4\
---------------------------------------------------------------------------
\4\ Each of the NITRD agencies supports diverse programs at
multiple scales. This description captures the dominant mode of each
agency.
---------------------------------------------------------------------------
The rich ecosystem of computing research approaches, collaborative
agencies and funding models has long made the U.S. the undisputed
leader in information technology, with concomitant benefits to our
national security, economic competitiveness and lifestyle.
4. Research Horizons and Risks: The Funding Monoculture
In a biological ecosystem, environmental changes or the death of a
species can change the ecosystem's set point or even lead to its death;
the NITRD ecosystem is no different. Today, the health of the NITRD
ecosystem is threatened, and the future of our national competitiveness
is at grave risk, due largely to an over-dependence on a single
research funding source, a single funding approach and inadequate
research funding overall. Through the 1990s, academic computing
research funding was dominated by two NITRD sources, DARPA and NSF,
with each filling complementary ecosystem niches based on different
project selection models, funding scales and assessment approaches.
From its inception, DARPA supported larger-scale, outcome-driven
initiatives and projects based on targeted solicitations. DARPA program
managers had broad latitude to assemble academic and industrial
consortia that built computing technology prototypes and transferred
promising prototypes into industry for commercialization. In a
complementary role, NSF funded exploratory, innovation-driven computing
research, funding peer-reviewed research proposals submitted by the
academic community. Although project funding levels were typically
lower than at DARPA, researchers were free to explore novel ideas of
their own choosing. NSF researchers not only filled niches not occupied
by DARPA, their most promising results often stimulated new DARPA
technology prototyping and transfer initiatives.
As an example, research flourished in computer architecture, system
software, programming models, algorithms and applications in the 1990s.
Computer vendors launched new initiatives, parallel computing startup
companies were born, and planning began for petascale systems, based on
integrated hardware, architecture, software and algorithms research.
This renaissance in parallel and high-performance computing research
was a direct consequence of the High-Performance Computing and
Communications (HPCC) program and interdependent agency initiatives,
notably DARPA and NSF. DARPA funded large-scale hardware prototypes and
software initiatives, while NSF supported exploratory research by
single investigators.
When DARPA shifted its funding and evaluation model to shorter-
term, ``go/no-go'' assessments and approaches, the ecosystem of funding
agencies and researchers reacted and adapted. Large-scale computing
research contracted, and those academic institutions and faculty who
has historically benefited from DARPA's largess turned to NSF for
research funding. This retreat of DARPA from funding fundamental
computing research at U.S. universities has left a hole in the overall
federal IT research ecosystem that other participating agencies have
been unable to fill. The types and scale of research changed and the
number of research proposals submitted to NSF rose precipitously, with
a concomitant decline in proposal success rates.
The National Science Foundation is now the predominate funder of
all academic computing research. Indeed, recent analyses show that NSF
provides 86 percent of all funding for academic computing research.\5\
The result is that NSF is now viewed by most academic researchers as
the only viable source of research funding. The notable exceptions are
the DOE SciDAC program and those faculty members who have strong ties
to the national laboratories.
---------------------------------------------------------------------------
\5\ National Science Foundation, Division of Science Resources
Statistics. 2008. Federal Funds for Research and Development: Fiscal
Years 2005-07. Forthcoming. Arlington, VA.
---------------------------------------------------------------------------
The consequences of this ecosystem shift are both deep and
profound, with several deleterious effects. First, fierce competition
for funding has made researchers risk averse. Today, those proposals
recommended for funding are far more likely to emphasize short-term,
incremental research that builds on well-understood approaches. Such
proposals are less controversial and more likely to win consensus
approval than those embodying high risk, ground-breaking ideas. This is
especially worrisome given the timeline of Figure 1, which shows the
long incubation period for these technologies between the time they
were conceived and first researched to the time they arrived in the
market as commercial products. In nearly every case, that lag time is
measured in decades.
Incremental advancement itself is not bad; it is the lifeblood of
the scientific process. However, just as a balanced retirement
portfolio includes an evolving mix of low risk, modest return
investments and higher risk, higher return investments, the long-term
success of our computing research ecosystem depends on a balance of
modest risk, moderate payoff research and higher risk, but high payoff,
revolutionary research. We must rebalance our research portfolio to
encourage greater innovation and risk taking.
Second, current academic structures necessitate research funding as
an external validation of quality and to sustain internal research
processes. Hence, faculty members face enormous institutional pressure
to seek external research funding for promotion, tenure and national
visibility. Because only a modest fraction of submitted proposals is
funded in many programs, faculty members now spend an inordinate
fraction of their time preparing, submitting and reviewing proposals.
It is not uncommon for an assistant professor to write five or even ten
proposals in a single year, hoping one or two will be funded. Hence, we
must address the funding shortfall that currently limits research
innovation.
5. Research Priority Areas: Identifying Innovation Foci
The seminal 1999 PITAC report, Information Technology Research:
Investing in Our Future,\6\ highlighted the importance of software
noting, ``Software is the new physical infrastructure of the
information age. It is fundamental to economic success, scientific and
technical research, and national security.'' The report also noted that
the diversity and sophistication of our software systems was growing
rapidly at a time we lacked the technologies to build reliable and
secure software systems and that even more perniciously, we were under-
investing in the research needed to develop those technologies. In
addition to the critical importance of software, the 1999 PITAC report
emphasized the importance of adequate research investment in scalable
information infrastructure and high-performance computing.
---------------------------------------------------------------------------
\6\ President's Information Technology Advisory Committee,
Information Technology Research: Investing in Our Future, http://
www.nitrd.gov/pitac/report, 1999
---------------------------------------------------------------------------
In 2007, PCAST revisited the 1999 PITAC technical priority areas,
concluding that the broad areas remained deeply relevant, albeit with
slight changes. Information technology systems that interact with the
physical world emerged as the new top priority--cyber-physical systems
where computing systems, sensors and actuators are deeply embedded in
engineered objects. Such systems are now both diverse and ubiquitous
and include our critical national infrastructure such as the electric
power grid, mobile and human-centered sensors (e.g., mobile biomedical
devices), environmental monitors and military systems. Such systems can
be difficult and costly to design, build, test, and maintain and the
consequences of failure can be catastrophic. However, the benefits are
enormous, including more efficient transportation systems, more
efficient energy generation and management and a reduced carbon
footprint for a diverse set of human activities.
One should rightly view systems that interact with the physical
world as a special case of the broader software priority identified by
PITAC. In this spirit, software remains the second broad priority
identified by PCAST, along with networking and digital data. The latter
two areas reflect the popularization of the Internet, with concomitant
challenges in security, scalability, resilience and management, and the
explosive growth of digital data, itself enabled by inexpensive sensors
and large-scale storage devices. Advances in these areas are also
essential to national security and to combating cyber crime. PCAST also
recognized the need for continuing emphasis on high-end computing,
cyber security and information assurance, human-computer interaction,
and information technology and the social sciences.
6. Workforce: Ensuring Quality and Quantity
In a knowledge economy, continued innovation and international
competitiveness depend on an adequate and continually renewed supply of
qualified and motivated workers. In the U.S., the IT workforce is
composed of those educated here--U.S. citizens, permanent residents and
international students--and the best and brightest from around the
world who choose to live and work here. We face both quantity
challenges, ensuring an adequate supply of IT workers, and quality
issues, creating curricula that match emerging technical trends and
that attract and excite sufficiently diverse cross-section of the
population. As the 2007 PCAST report noted,
Although the overall supply of networking and information
technology specialists is expected to grow in response to the
growth in total demand, at current rates of enrollment and
graduation, shortfalls in the numbers of highly qualified
computer scientists and engineers graduated at the
undergraduate and doctoral levels are likely. Women and other
under-represented groups will constitute a declining proportion
of the new graduates.
The stereotype of a geek who writes code in a small cubicle and who
eschews human interaction is neither reflective of the diversity of
modern computing and nor of computing's role in all aspects of society,
from the arts and humanities through business practice to science and
engineering. Many academic, federal and private groups are working
assiduously to dispel this stereotype and raise the image of computing
among potential students. The Image of Computing Task Force\7\ was
created by a consortium of companies and computing professional
societies to ``expose a realistic view of opportunities in computing''
and to ``educate the public and those with the aptitude and interest to
pursue computing careers, on the increasing vital role computing plays
in every major field.'' In addition, the CRA Committee on the Status of
Women (CRA-W),\8\ the National Center for Women and Information
Technology (NCWIT) \9\ and the Coalition to Diversity Computing (CDC)
\10\ are all highlighting the importance of diversity in computing and
the opportunities for creative and engaging careers.
---------------------------------------------------------------------------
\7\ Image of Computing, http://www.imageofcomputing.com
\8\ CRA Committee on the Status of Women, http://www.cra.org/
Activities/craw
\9\ National Center for Women and Information Technology, http://
www.ncwit.org
\10\ Coalition to Diversity Computing, http://www.cdc-computing.org
---------------------------------------------------------------------------
In addition to increasing awareness of information technology as a
vibrant, attractive and relevant problem-solving skill in the 21st
century knowledge economy, computing professional societies and
universities are working to revamp curricula that have changed
relatively little since the 1970s. The changes include increasing
multi-disciplinary computing education (i.e., computing and its
applications to another discipline), multi-track curricula that allow
students to create degree programs that better match their interests
and emphasizing the power of computing as a general-purpose problem
solving tool. As a complement to image education and curricula reform,
PCAST also recommended increasing the number of multi-year graduate
fellowships offered to U.S. students.
These image, curricula and fellowship reforms potentially address
the shortfall of domestic students. However, the U.S. information
technology ecosystem has long been a magnet for talented students,
researchers and workers from around the world. Such individuals
increasingly find attractive educational, research and professional
opportunities in their home countries. It is in the best interests of
the U.S. to retain the best and brightest international students who
obtain graduate degrees in this country, often supported by research
grants and contracts. Hence, PCAST also recommended streamlining the
process for obtaining visas for non-U.S. students admitted to
accredited graduate degree programs and to make it routine for foreign
nationals who have obtained advanced degrees in NIT subjects at
accredited U.S. universities to be permitted to work and gain
citizenship by easing visa and permanent resident processes for them.
7. NITRD Coordination: Strategic Planning and Execution
Without doubt, the NITRD program has been effective in fostering
informal communication and coordination across agencies, both
collectively and via the National Coordination Office (NCO). The NCO
annually solicits and reports agency spending on NITRD Program
Component Areas (PCAs). Though each federal agency is represented
within a NITRD Interagency Working Group (IWG) on IT research and
development, the IWG has no budget authority over any of the
participating agencies or the PCAs, nor does the NCO. Each agency
controls its own budget and sets its own goals exclusively on the
perceived appropriateness of that funding to the agency's mission.
In practical terms, this means the IWG function in NITRD is largely
one of information sharing among agency representatives on what the
agencies plan to do and have done. Although the resulting NCO data is
useful, it is a retrospective view of agency decisions and priorities,
rather than an assessment of program priorities and progress against
those plans. The process also tends to bias the process toward
incremental, agency-specific agendas, making the NITRD program less
effective in managing larger-scale, coordinated projects than span
multiple agencies.
In a globally competitive world, we must plan more strategically
and increase agency accountability for execution against those
strategic plans. This will require greater interagency coordination and
collaboration across PCAs to facilitate research and development
transition within and across agencies, both to support fundamental
research and to enable larger, multi-agency projects.
8. Remaining Competitive: A Call to Action
To maintain the health and vibrancy of the U.S. information
technology ecosystem, we must fully fund the agencies and programs
included in the America COMPETES Act. I commend you and your
colleagues, Mr. Chairman, for working so hard for its passage last
year. It sent a powerful signal about the importance of the federal
role in supporting fundamental research in the physical sciences,
including information technology. I also appreciate your efforts to see
the promise of the COMPETES Act realized in appropriations. The funding
authorized in the Act would help drive the core of the IT innovation
engine--the fundamental information technology research in U.S.
universities and national laboratories supported by the National
Science Foundation, the Department of Energy's Office of Science, and
the National Institute of Standards and Technology.
The focus within the COMPETES Act on programs that aim to increase
the number of students who enter science, technology, engineering and
mathematics (STEM) fields is also crucial to the future of information
technology research. As I noted earlier, the projected demand for IT
professionals over the next 10 years--positions that require at least a
Bachelor's degree in computer science or computer engineering--exceeds
all other science and engineering disciplines combined. Encouraging
U.S. students to enter the science and engineering education pipeline,
as is the focus of many of the programs included in the COMPETES Act,
will help ensure that those projected workforce needs are addressed.
The many provisions in the Act that seek to increase the participation
of women and minorities in science and engineering fields--two
populations that are woefully under-represented in computing--are
especially important.
Adequate funding is critically important, but it is not sufficient;
this funding must be invested wisely in our information technology
ecosystem. The unilateral decision by any agency to change the
direction, scope and mechanisms for its research investments has
consequences across the entire NITRD ecosystem--federal agencies,
universities and industry. Such changes must not be undertaken without
due consultation and consideration of broad consequences. We must
rebalance agency participation in the NITRD program so that the crucial
responsibility of supporting fundamental research in computing is not
borne solely by one agency.
We must also create an interagency IT research and development
strategic plan, complemented by a roadmap and a set of associated
metrics that define interagency expectations and accountabilities. An
evolving, strategic vision of information technology, together with an
appropriate balance of short-range, low risk and long-range, high risk
projects is essential if we are to remain global leaders. The 1999
PITAC report recommended creation of large-scale Expeditions to the
21st Century, revolutionary expeditions whose mission
. . . will be to report back to the Nation what could be
accomplished by using technologies that are quantitatively and
qualitatively more powerful than those available today. In
essence, these centers will create ``time machines'' to enable
the early exploration of technologies that would otherwise be
beyond reach for many years.
We would do well to embrace this vision and recommendation,
ensuring that we fund a mix of projects, large and small, low and high
risk, and both short- and long-term.
Finally, we must also have appropriate oversight and review of our
research investment and accountability against strategic plans. The
President's Information Technology Advisory Committee (PITAC) was
authorized by Congress as a federal advisory committee under the High-
Performance Computing Act of 1991 and the Next Generation Internet Act
of 1998, with responsibility to assess advanced information technology
and review the NITRD program. PITAC functioned as a separate
Presidential advisory committee until its roles and responsibilities
were assigned by Executive Order in 2005 to the President's Council of
Advisors on Science and Technology (PCAST).
PCAST has a broad scope that spans all of science and technology, a
challenging and important portfolio. Given the importance of IT
research and technology to our nation's economy, national security,
military readiness and research enterprise, an independent PITAC is
needed that can devote the time, energy and diligence to ongoing
assessment of successes, challenges, needs and opportunities in
information technology. I base this opinion on my own experience as a
previous member of PITAC and a current member of PCAST. Simply put, the
NITRD program is best served by a stand-alone PITAC composed of
computing experts from academia and industry.
In summary, information technology is a universal intellectual
amplifier, advancing all of science and engineering, powering the
knowledge economy, enhancing the quality of our health care, and
transforming how we work, play and communicate. With vision, strategic
investment and coordination, the U.S. NITRD program can and will
continue to be the world's leader.
Mr. Chairman, thank you and this committee for your interest in the
future of the NITRD program and its importance to U.S. competitiveness.
Thank you very much for your time and attention. At the appropriate
time, I would be pleased to answer any questions you might have.
Biography for Daniel A. Reed
Daniel A. Reed is Director of Scalable and Multicore Computing
Strategy at Microsoft. Previously, he was the Chancellor's Eminent
Professor at the University of North Carolina at Chapel Hill, as well
as the Director of the Renaissance Computing Institute (RENCI), which
explored the interactions of computing technology with the sciences,
arts and humanities. He formerly held the Edward William and Jane Marr
Gutgsell Professorship at the University of Illinois at Urbana-
Champaign, where he was Professor and Head of the Department of
Computer Science and Director of the National Center for Supercomputing
Applications (NCSA). At Illinois, he also led National Computational
Science Alliance, a consortium of roughly fifty academic institutions
and national laboratories to develop next-generation software
infrastructure of scientific computing. He was also one of the
principal investigators and chief architect for the NSF TeraGrid.
Dr. Reed is a member of President Bush's Council of Advisors on
Science and Technology (PCAST) and a former member of the President's
Information Technology Advisory Committee (PITAC). He recently chaired
a review of the federal networking and IT research (NITRD) portfolio,
and he is Chair of the Board of Directors of the Computing Research
Association (CRA), which represents the research interests of
universities, government laboratories and industry. He received his
Ph.D. in computer science in 1983 from Purdue University.
Chairman Gordon. Thank you, Mr. Reed, and also thank you
for pointing out the need to continue the funding for COMPETES.
I think sometimes the real world out there doesn't understand
the difference between authorization and appropriation, and we
have got to continue to move forward. And thank you for
bringing that up.
I will also point out that in the COMPETES Act, we did
change the review to two years, and we made it a stand-alone
committee also.
Dr. Stewart, we would love to hear from you.
STATMENT OF DR. CRAIG A. STEWART, CHAIR, COALITION FOR ACADEMIC
SCIENTIFIC COMPUTING; ASSOCIATE DEAN, RESEARCH TECHNOLOGIES,
INDIANA UNIVERSITY
Dr. Stewart. Let me begin by thanking Chairman Gordon,
Ranking Member Mr. Hall, Messrs. Hills and Carson of Indiana,
and all Members of the House Science and Technology Committee
for the opportunity to be here today.
I am Chair of the Coalition for Academic Scientific
Computation, or CASC. I am offering testimony as requested by
Chairman Gordon regarding the President's Council of Advisors
on Science and Technology 2007 Report, Leadership Under
Challenge, Information Technology Research and Development in a
Competitive World. To provide context for this testimony, CASC
is an educational, non-profit organization dedicated to using
advanced computing technology to accelerate scientific
discovery for national competitiveness, global security, and
economic success.
There are a total of 53 CASC members, colleges, and
universities, and research labs in 36 states and the District
of Columbia. I note that Members of the Committee represent a
total of 24 states, 19 of which are home to at least one CASC
member.
As stated in the PCAST Report, we must improve the
networking and information technology ecosystem in the United
States to maintain and extend our competitive advantage and
innovation. The NITRD Program support of 13 federal agencies
including DOD, DOE, DARPA, NASA, NIH, NIST, and NSF has
accelerated information technology innovation and led to new
insights in science, technology, and medicine. These advances
have led to valuable changes in the private sector as we all
know.
CASC fully supports the overall recommendations of the 2007
PCAST Report. The recommendations in that report, if well-
supported by finding and executed aggressively, will contribute
greatly to continued U.S. leadership in networking and
information technology.
Without overarching endorsement as the key point of this
testimony, CASC would like to make a few suggestions to
emphasize and add to the PCAST recommendations. First, federal
investment in NIT research and development will be most
valuable in the long run if investment patterns in the many sub
areas included within NITRD are as consistent as possible over
time. The PCAST Report makes several important recommendations
regarding workforce development. We agree with these
recommendations and would like to suggest additional areas of
emphasis. Programs that will increase the number of students
who choose a major related to NIT after entering college
undecided on a major and continue to strengthen and expand the
emphasis on science, technology, engineering, and mathematics
disciplines in primary and secondary education.
We commend Chairman Gordon and the Members of the Committee
as a whole for leadership in creating and supporting the
development of the STEM program. We hope you might consider
expanding it to include greater emphasis on computing in the
future.
CASC would also like to expand on the report's
recommendation regarding a strategic roadmap for federal
investments in high-end computing research and development. In
addition to the recommendations made in the report, such a plan
should implement methods for sustained support and maintenance
of software critical to the U.S. networking and information
technology agenda. This plan should also support the
coordination of U.S. high-end computing facilities in a way
that maximizes the total benefit to U.S. national interest by
leveraging investments at the college, university, State, and
regional levels in addition to federal investments.
In closing, let me return to the title of the 2007 PCAST
Report, Leadership Under Challenge. U.S. leadership is indeed
under challenge in many ways across the globe. As regards,
networking information technology, the current challenges are
without precedent. Without strong investment, the United States
is at risk of losing its longstanding position of global
leadership in networking information technology. The
consequences of that would be catastrophic. However, the
recommendations made in the PCAST Report if enacted and well-
funded, will continue and extend U.S. leadership in networking
information technology and fuel future U.S. global leadership
and innovation generally. This will lead to continued and
improved prosperity, health, and security for Americans and
indeed all citizens of the world.
Thank you again for the opportunity to appear before you
today. I should note that my testimony this morning has been
endorsed by a formal vote of CASC members. One CASC member's
voting representative was unavailable due to travel. The
remaining 52 have voted unanimously to endorse my testimony
this morning. I hope that these remarks have been helpful to
the Committee. I am happy to answer any questions now or at any
time in the future.
[The prepared statement of Dr. Stewart follows:]
Prepared Statement of Craig A. Stewart
1. Background and context
I am pleased to have this opportunity to provide testimony to the
House Science and Technology Committee in response to a request from
Chairman Gordon. Chairman Gordon, in his letter of invitation to the
Coalition for Advanced Scientific Computing, asked for comments on the
President's Council of Advisors on Science and Technology (PCAST) 2007
report Leadership Under Challenge: Information Technology R&D in a
Competitive World\1\ and the merit of the recommendations therein. To
provide context for this testimony, I serve as the chair of the
Coalition for Advanced Scientific Computing (CASC) (http://
www.casc.org), an educational nonprofit 501(c)(3) organization with 53
member institutions, representing many of the Nation's most forward
thinking universities and computing centers. CASC is dedicated to
advocating the use and development of the most advanced computing
technology to accelerate scientific discovery for national
competitiveness, global security, and economic success, as well as to
developing a diverse and highly skilled 21st century workforce. My
testimony this morning has been endorsed by a majority of the members
of CASC, and represents the general consensus of opinion within CASC.
---------------------------------------------------------------------------
\1\ President's Council of Advisors on Science and Technology
(PCAST). 2007. Leadership Under Challenge: Information Technology R&D
in a[0] Competitive World. http://www.nitrd.gov/pcast/reports/PCAST-
NIT-FINAL.pdf
---------------------------------------------------------------------------
I also serve Indiana University as the Associate Dean for Research
Technologies and the Chief Operating Officer for the Pervasive
Technology Labs at Indiana University. As such, I am responsible for
many of the advanced networking and information technology services
provided to Indiana University researchers. Through support from the
State of Indiana and Federal agencies, I am also responsible for
services delivered to public and private sector researchers in Indiana
and researchers at institutions of higher education throughout the U.S.
I came to be involved in networking and information technology
originally as a biologist. I thus value advanced technology first and
foremost for what it can do practically to improve the quality of human
life and our understanding of the world around us.
2. Key observations
In their letter submitting the 2007 PCAST report, Co-Chairs John H.
Marburger III and E. Floyd Kvamme summarized in two sentences the
challenge facing the U.S. in networking and information technology
(NIT):
``While the United States clearly is the global NIT leader
today, we face aggressive challenges from a growing list of
competitors. To maintain--and extend--the Nation's competitive
advantages, we must further improve the U.S. NIT ecosystem--the
fabric made up of high-quality research and education
institutions, an entrepreneurial culture, strong capital
markets, commercialization pathways, and a skilled NIT
workforce that fuels our technological leadership.''
CASC strongly endorses this statement and the findings and
recommendations included in the report. The key summary of the past,
included on page 1, that ``. . .. the NITRD [Networking and Information
Technology Research and Development] Program has by and large been
effective at meeting agency and national needs'' is correct. Indeed,
the NITRD program's support of fourteen Federal agencies, including the
Department of Defense, Department of Energy, and DARPA, has accelerated
innovation in information technology, leading to new insights and
practical, valuable changes in industry (including improved fuel
efficiency, health and medical care, homeland security, and the
creation of many physical devices that improve our productivity and
overall quality of life).
The Community I represent fully supports the overall
recommendations stated in the PCAST report. General George S. Patton
stated, ``A good plan, violently executed now, is better than a perfect
plan next week.'' The findings of PCAST are--overall--spot on. It is
easy to quibble over details, but in general the recommendations, if
executed aggressively, would be far better than inaction or
continuation with the status quo in the NITRD Program.
With that overarching endorsement as the key point of this
testimony, we would like to make three additional points to emphasize
and add to the PCAST recommendations regarding investment patterns over
time, workforce development, and creation and implementation of a High
End Computing Research and Development plan.
3. Pattern of investment over time
Without strong, continued, and consistent investment in networking
and information technology (NIT), the U.S. will not have the
administrative and technical leadership to support consistent and
directed change. Government investment in NIT will be of greatest value
if there is consistency in levels of investment over time. The men and
women who execute the national NIT agenda represent a tremendous store
of experience, skill, and knowledge. The uniform experience of CASC
members is that when there are strong variations in funding in specific
areas of NIT over time, lean times for particular areas of research in
NIT cause skilled professionals to leave public sector NIT research.
This means that years of investment by the government in developing a
knowledge and experience base in individuals who desire to pursue a
career in the public service sector are lost to the public sector, not
to return even when funding for particular areas is subsequently
restored. U.S. global competitiveness, innovation, and homeland
security are thus best served by consistent and strong investment in
basic NIT research; advanced NIT facilities to support advanced
research and development in science, engineering, and technology; and
research in developing and delivering the next generation of such
advanced NIT facilities.
4. Workforce Development
The PCAST report makes several important recommendations regarding
workforce development aimed at increasing the supply of professionals
with Bachelor's, Master's, and doctoral degrees in NIT areas. The
recommendations focus on actions that should increase the supply of
skilled NIT professionals in the U.S. in the short-term. This is
critically important, and CASC supports all of those recommendations.
We would like to make two suggestions for funding emphasis that are in
addition to the recommendations made in the report.
Recommendation: Increase the number of students receiving a
Bachelor's degree in a field related to NIT by funding programs that
encourage students to explore NIT majors. An effective way to do this
would be to support programs that use tele-collaboration technologies
to enhance the NIT-related course offerings at small colleges and
universities, particularly those that serve large populations of
students from groups traditionally under-represented among NIT
professionals. For example, students at Jackson State University, an
HBCU (Historically Black College or University), and Navajo Technical
College (a college located within the Navajo Nation) took, via
teleconference, computer science courses from IU School of Informatics
Professor Geoffrey C. Fox. Students who took these courses indicated
that they found the classes inspirational and that they affected their
career plans. This activity was enabled by relatively modest funding
from the National Science Foundation. Similarly, Thomas Sterling, the
inventor of Beowulf computing and a computer science professor at
Louisiana State, has taught classes in high performance computing
classes via tele-collaboration to students of the University of
Arkansas and Louisiana Tech. Increased investment in collaborative
distance education, either in absolute terms or as a relative share of
the NITRD budget, would have disproportionately great long-term impact
on the supply of professionals with college degrees in NIT.
Recommendation: Continue to strengthen and expand the emphasis on
STEM (Science, Technology, Engineering, and Mathematics) disciplines in
elementary and secondary education, so as to increase the absolute
numbers and relative percentages of high school graduates who plan to
enter college in an NIT-related discipline. We would like to commend
Chairman Gordon for his leadership in creating and supporting the
development of the STEM program. The uniform experience of CASC member
organizations is that within their home states, there are areas where
the educational system and social environment do not provide adequate
incentive or opportunity for our young people to become excited by STEM
disciplines and then acquire the primary and secondary education needed
to successfully pursue an undergraduate (and then advanced) education
in NIT-related areas. The PCAST report recommends steps to increase the
importing of talent to the U.S. from abroad at the same time that we
are losing the opportunity to develop our own talent. Each CASC
institution can provide data to support this. In my home State of
Indiana, for example innately bright young people in the rural
southwest and urban northwest of the state are lost to the U.S. 21st
century workforce because they are provided neither the inspiration nor
the education that would enable them to pursue careers in NIT. We
recognize that this area is beyond the statutory responsibility of
NITRD, but it is important and related to NITRD and the PCAST
recommendations. Chairman Gordon, we hope that you might now consider
leveraging the successful STEM program by expanding it to include
Computing.
5. High End Computing Research and Development Roadmap
The PCAST report makes several recommendations regarding
investments in High End Computing. We endorse those recommendations and
would like to expand on one of the recommendations (made on page 40 of
the PCAST report):
``Recommendation: The NITRD Subcommittee should develop,
implement, and maintain a strategic plan for Federal
investments in HEC [high-end computing] R&D, infrastructure,
applications, and education and training. Based on the
strategic plan, the NITRD Subcommittee should involve experts
from academia and industry to develop and maintain a HEC R&D
roadmap.''
As noted in the PCAST report, such a roadmap should be based on the
2004 Federal Plan for High-End Computing.\2\ Since the writing of that
2004 report, several new developments in the NIT ecosystem have taken
place, creating new opportunities for increased innovation, more
widespread practical benefits resulting from those innovations, and
enhanced leverage of federal investments. CASC offers two suggestions
regarding the plan called for in this recommendation, to be added to
the bullet points listed on page 40 of the PCAST report. A strategic
plan for federal developments in HEC R&D should:
---------------------------------------------------------------------------
\2\ National Science and Technology Council. Federal Plan for High-
End Computing. Washington, D.C.: May 2004, available at http://
www.nitrd.gov/pubs/2004-hecrtf/
20040702-hecrtf.pdf
Implement methods for sustainable support for
software development critical to the U.S. NIT agenda. This must
include supporting creation of complexity-hiding interfaces
that will dramatically expand the ability of scientists and
engineers generally to leverage and effectively use HEC
---------------------------------------------------------------------------
infrastructure.
Support the coordination of U.S. cyberinfrastructure
that maximizes the total benefit to U.S. national interests by
taking best advantage of investments at the college,
university, State, and regional levels, in addition to federal
investments.
I would like to briefly explain these points below.
Implement methods for sustainable support for software development
critical to the U.S. NIT agenda. This must include supporting creation
of complexity-hiding interfaces that will dramatically expand the
ability of scientists and engineers generally to leverage and
effectively use HEC infrastructure. The Federal Government needs to
significantly increase its investment in research, development, and
sustained support of important software tools. As noted in the PCAST
report, software critically important to U.S. global competitiveness is
not always viable as a commercial product, yet sustaining it over time
is critical to U.S. interests. Sometimes open source software
development is a solution. A new approach--community source software--
is emerging within universities to coordinate and leverage efforts in
development of educational and financial management software. This
approach may or may not be applicable to scientific software. But it is
notable that a relatively modest investment by the Mellon Foundation
enabled the Sakai Collaboration\3\ to develop a completely new approach
to sustainability of educational software. Similarly a modest
investment by the William and Flora Hewlett Foundation enabled the
Connexions\4\ project to develop a global open and free repository for
authors, instructors, and students to share and develop educational
material. CASC recommends that the Federal Government investigate and
support new models for scientific software sustainability in addition
to those already in use.
---------------------------------------------------------------------------
\3\ http://sakaiproject.org/
\4\ http://cnx.org/
---------------------------------------------------------------------------
An important new trend in HEC software environments is the concept
of a Science Gateway. A Science Gateway is a web-accessible tool that
provides end-to-end support for a scientific work flow, such as the
prediction of tornadoes or the analysis of an earthquake or a genome.
For example, one Science Gateway developed with NSF support provides an
intuitive interface that allows a weather expert to select input data
from Doppler radars, process multiple predictions of tornado formation
using some of the U.S.'s fastest supercomputers, and produce a
visualization on a laptop computer in time to send emergency warnings
and save lives. Science Gateways provide this sort of sophisticated
capability to scientists and engineers without requiring that such
people, who have invested years in becoming experts in their own
specific disciplines, also invest years in becoming expert
computational scientists. Using HEC systems to predict tornadoes,
analyze genomes, understand earthquakes, etc., should be as easy--for
researchers who understand the underlying science--as buying a book
over the Internet; identifying and understanding the critical aspects
of terabytes of data should be like starting with a web-accessible
image of North America and zooming in on your own back yard. For
decades, national and discipline-specific agendas of a few grand
challenge problems in high end computing have catalyzed innovation
within the U.S. Today there are thousands of important theoretical and
practical problems that can and will be solved if the HEC
infrastructure of the U.S. can be made more easily usable. In addition,
such complexity-hiding interfaces give undergraduate and even high
school students the opportunity to use high-end computing, which will
aid the STEM education and 21st century workforce development I have
already recommended.
Support for development of complexity-hiding interfaces must be in
addition to the much-needed investments in software development on
which such gateways depend and which are already called for in the
PCAST report. For example, new programming models and approaches to
programming are needed to take advantage of emerging HEC architectures,
particularly multi-core processors and specialized computational
hardware. In addition, today's high quality (including 3D) computer
displays, enhanced by research and development in visualization, can
provide new tools for extracting insight from the massive streams of
data now produced by digital instruments.
Support the coordination of U.S. cyberinfrastructure that maximizes
the total benefit to U.S. national interests by taking best advantage
of investments at the college, university, State, and regional levels,
in addition to Federal investments. While the term cyberinfrastructure
is not used in the PCAST 2007 report, it is useful in a discussion of
NIT and national competitiveness. The first usage of the term
cyberinfrastructure that I can find is from a 1998 press briefing by
Richard Clarke, then National Coordinator for Security, Infrastructure
Protection, and Counter-terrorism.\5\ The term became widely used after
its inclusion in a very important report by a blue-ribbon committee
commissioned by the NSF.\6\ There are several definitions of
cyberinfrastructure; the one I like best (admittedly developed by my
group at Indiana University) is as follows:
---------------------------------------------------------------------------
\5\ Press briefing by Richard Clarke, National Coordinator for
Security, Infrastructure Protection, and Counter-Terrorism;; and
Jeffrey Hunker, Director of the Critical Infrastructure Assurance
Office. 22 May, 1998. http://www.fas.org/irp/news/1998/05/980522-
wh3.htm
\6\ Report of the National Science Foundation Blue-Ribbon Advisory
Panel on Cyberinfrastructure. http://www.nsf.gov/od/oci/reports/
atkins.pdf
``Cyberinfrastructure consists of computing systems, data
storage systems, advanced instruments and data repositories,
visualization environments, and people, all linked together by
software and high performance networks to improve research
productivity and enable breakthroughs not otherwise possible.''
\7\
---------------------------------------------------------------------------
\7\ Indiana University Cyberinfrastructure Newsletter, March, 2007.
http://racinfo.indiana.edu/newsletter/archives/2007-03.shtml
Cyberinfrastructure is indeed the foundation for innovation for our
nation. Leadership class systems within the national
cyberinfrastructure are funded by NITRD, and that is likely to continue
for some time. However, the broad foundation for innovation will best
serve the needs of the Nation if Federal leadership can aid the
coordination of the collective cyberinfrastructure assets funded by
NITRD agencies and those funded by other sources, including colleges,
universities, states, and regional consortia. The resulting extension
and leverage of Federal investment in NIT, HEC, and cyberinfrastructure
would be tremendous and far-reaching, enabling the U.S. to increase its
global competitiveness far beyond what would be capable on the basis of
---------------------------------------------------------------------------
federal investment without such coordinated leverage.
6. Conclusion
In conclusion, let me return to the starting point of the PCAST
report. NITRD has been tremendously important to U.S. innovation and
global competitiveness, the quality of life of Americans, and the
security of our homeland. CASC members endorse the recommendations
contained in the PCAST report, and hope that the comments made in this
testimony regarding particular areas of emphasis or addition of
recommendations will be of value to this Committee as it embarks upon
activities to plan for an even better future of new, important, and
practical accomplishments through legislation related to NITRD.
The 2007 PCAST report is titled Leadership Under Challenge:
Information Technology R&D in a Competitive World. U.S. leadership is
indeed under challenge in many ways across the globe. As regards
networking and information technology, these challenges are
unprecedented. Without strong investment in NIT, the U.S. is at risk of
losing its longstanding position of global leadership, and the
consequences of this would be catastrophic. However, the
recommendations made in the PCAST report, if enacted into legislation
and well funded, will continue and extend U.S. leadership in network
and information technology, and will fuel future U.S. global leadership
in innovation. This will lead to continued and improved prosperity,
health, and security for Americans and indeed all citizens of the
world.
Thank you for the opportunity to appear before you today. I am
happy to answer any questions now or at any time in the future.
Biography for Craig A. Stewart
Craig Stewart is Associate Dean for Research Technologies and Chief
Operating Officer for the Pervasive Technology Labs at Indiana
University. In these roles, Dr. Stewart oversees activities conducive
to and supporting research in advanced information technology. He
received his Ph.D. in Biology from Indiana University in 1988, and has
held a variety of positions in Information Technology at Indiana
University. His longstanding career interests are in high performance
computing and computational biology. In high performance computing his
areas of concentration are HPC architectures and grid computing. In the
area of computational biology his areas of concentration are
computational phylogenetics, computationally intensive simulation
methods in systems biology, and biomedical data grids. Dr. Stewart is
currently chair of the Coalition for Academic Scientific Computing.
Dr. Stewart served as guest editor for Bioinformatics: transforming
biomedical research and medical care, the November 2004 special issue
of Communications of the Association for Computing Machinery. He has
co-authored numerous papers, including Measuring quality, cost, and
value of IT services in higher education for the 2001 American Quality
Congress, Parallel computing in biomedical research and the search for
petascale biomedical applications for Advances in Parallel Computing in
2004, and Implementation of a distributed architecture for managing
collection and dissemination of data for fetal alcohol spectrum
disorders research for Grid Computing in Computational Biology in 2006.
Dr. Stewart has also presented many tutorials, including a 2005
introduction to computational biology at High Performance Computing
Center, Stuttgart, Germany. He also helped lead two winning projects at
the premier annual international supercomputing conference: Global
Analysis of Arthropod Evolution, the 2003 HPC Challenge winner; and
Using the Data Capacitor for Remote Data Collection, Analysis, and
Visualization, the 2007 Bandwidth Challenge winner.
Dr. Stewart is an active participant in several federally funded
grants, including: TeraGrid Resource Partners (NSF); Acquisition of
PolarGrid: Cyberinfrastructure for Polar Science (NSF); the Open
Science Grid (NSF/NIH); and Major Research Infrastructure: Data
Capacitor (NSF).
Chairman Gordon. Thank you, Mr. Stewart. As I mentioned
earlier, we are in the process of trying to gain more
information. The Academy will be an important part of that
information. As you know, Baron Hill is fortunately your Member
of Congress, and so he will be taking a direct row on this and
we want you to be a conduit for information for the Academy and
for Baron to play a role in it. Thank you very much.
And finally, Mr. Don Winter.
STATEMENT OF MR. DON C. WINTER, VICE PRESIDENT, ENGINEERING AND
INFORMATION TECHNOLOGY, PHANTOM WORKS, THE BOEING COMPANY
Mr. Winter. Good morning, Mr. Chairman, Ranking Member Hall
and Members of the Committee. I am Don Winter, Vice President
of Engineering and Information Technology at Boeing Phantom
Works. I am grateful for the invitation to speak with you on
the NITRD Program, specifically those focused on cyber-physical
systems.
I was impressed with the way in which these recommendations
were developed, bringing stakeholders from government,
academia, and industry together with a common focus on national
competitiveness.
PCAST Report builds a sound case for the recommended
research focus areas, including the area of specific interest
to me, the cyber-physical systems.
The subject of research on cyber-physical systems or CPS is
of great importance to the aerospace industry as a whole and to
our nation. The use of CPS is increasing, their complexity is
growing at an exponential rate. Demands for higher performance
and lower cost for commercial and military systems are driving
next-generation systems to be highly networked and highly
dynamic in nature. Moreover, systems will need to be designed
to exhibit robust and predictably safe behavior in these highly
dynamic environments. Future aerospace systems will require
cyber-physical systems of even greater complexity. Systems will
operate with high degrees of autonomy or collaborate among
themselves to achieve dramatic gains in operational
effectiveness. New cyber-physical system attributes such as
active resource management, dynamic scheduling, and software
enabled control mode changes will be needed to support these
behaviors. These emerging challenges call for cyber-physical
systems on a grand scale. Research that addresses validation
and verification of the complex interactions between system
modules is critical. Without advances in these technologies,
the costs and risks of developing next generation cyber-
physical systems of this scale may be prohibitive and have a
significant impact on the industry.
Many of our systems are safety-critical and require
certification by the FAA or equivalent military authority. Many
of our military systems will need to support coalition
operations with multi-level security requirements. Our systems
must also be hardened to withstand a future cyber attack.
Because of these unique requirements and the relatively small
number of end systems, we do not expect to see large investment
from the commercial IT sector in these technologies. In order
to achieve these cross-cutting capabilities, we will need
advances in design technologies such as model-based development
tools and validation environments to build systems rapidly and
affordably. Moreover, we will require research and product
focus technology in software reuse, real-time theory,
languages, and product line CPS architectures. It can be
applied to many different end systems.
We have achieved some measure of progress. Over the past 10
years, Boeing has developed metal-ware based product line, CPS
architectures, notably the bold stroke architecture for
tactical aircraft avionics and the system of systems common
operating environment are SOSCO for the future combat systems.
To support our military system developments, and substantial
gains in productivity were realized.
What is the way ahead? Efforts today that have been
fragmented across industry and limited by internal funding
constraints.
CPS investments cross multiple technology domains will
require an industry level critical mass to achieve the needed
result. Other industries, notably automotive, energy management
and control, and medical face similar CPS trends and pressures
and have expressed their desire to participate. WE need a
national strategy in which long-term CPS technology needs are
addressed by combined government and corporate investment.
Boeing for its part can focus long-term CPS investments of
collaborative research in which we provide the challenge
problems and in-kind participation and government industry
research consortia. Although I don't speak for them, I am
confident my industry partners are willing to do the same. We
also need to develop new ways to facilitate the transition of
research products back into industry and into our products.
The point is critical, and again, as a matter of national
competitiveness, The European Union's Advanced Research and
Technology for Embedded Intelligence Systems, ARTEMIS, program
is funded by a public-private investment of over $7 billion and
is persuading R&D to achieve ``world leadership in intelligent
electronic systems'' by 2016. European industry is fully
partnered with government and academia in ARTEMIS. From our
perspective, an active partnership of this nature in CPS is
essential to reap the benefits of this advanced research. This
partnership needs to reach deeper than the arm's-length
approach used for industry involvement today.
In summary, we support the proposed expansion of the NITRD
program's research objectives to address cyber-physical systems
and we look forward to the opportunity to participate. That
concludes my testimony. I would be pleased to respond to your
questions.
[The prepared statement of Mr. Winter follows:]
Prepared Statement of Don C. Winter
Good morning, Mr. Chairman, Ranking Member Hall and Members of the
Committee.
I am Don Winter--Vice President of Engineering and Information
Technology at Boeing Phantom Works. I am grateful for the invitation to
speak with you on this subject of research on cyber-physical systems
(CPS), a topic of great importance to the Boeing Company, the aerospace
industry as a whole, and to our Nation. I have a great interest in this
subject because of my current position managing an annual R&D budget of
over $300M and my past position as one of the founders of the Bold
Stroke R&D initiative at Boeing, focused largely on advancing the state
of the art in cyber-physical systems.
Boeing has a somewhat unique perspective on cyber-physical systems
due to our prominent position in both the military and commercial
aerospace markets. Cyber-physical systems are pervasive at Boeing, and
in the aerospace industry at large. They are becoming increasingly
prevalent in other sectors, notably automotive and energy management.
Their importance to our products is huge and their complexity is
growing at an exponential rate. Demands for higher system performance
and lower system cost for commercial and military systems are driving
next generation systems to be highly networked and highly dynamic in
nature. Lower recurring and maintenance costs will be derived from
integrated vehicle health management that enhances system reliability
and reduces logistics and maintenance costs. Moreover, systems will
need to be designed to exhibit ``predictably safe'' behaviors in an
uncertain environment.
In the 70's and 80's aerodynamics and structures accounted for
nearly 90 percent of the development cost of a transport aircraft, with
cyber-physical system development accounting for less than 10 percent.
The trend has reversed, and cyber-physical system design, development,
validation and certification account for nearly half of development
costs for current generation system, and for next generation systems
this percentage is expected to rise to 50 percent or more.
Several examples are germane and illustrate the exponential growth
in software and system complexity of our modern systems. The 747-400
first flew in the late 1980's. The size of the software for the on-
board cyber-physical systems is on the order of 10MB. The Boeing 777
first flew in the early 1990's. Its flight software size is an order of
magnitude larger--100MB (on the order of 10 million SLOC). As we evolve
to systems such as 787, software size and system complexity will be
increased by two or more orders of magnitude.
These are cyber-physical systems on a grand scale. Research that
can support validation and verification of the complex interactions
between system modules is highly important. Without advances in these
technologies, the cost and risk of developing next generations of
cyber-physical systems of this scale may be prohibitive, and have a
significant impact on the aerospace industry.
The trends towards CPS complexity are not exclusive to the
aerospace industry. The automotive industry has a similar experience.
For the last several years, Boeing has been participating in CPS forums
across aerospace, automotive, and energy sectors. At a May 2007 CPS
Roundtable, representatives from USCAR (the U.S. Council for Automotive
Research--an umbrella organization for collaborative research among
Chrysler, Ford, and General Motors) reported similar trends. Currently
the percentage of vehicle cost due to electronics content is
approximately 30 percent. The electronics content is increasing in
complexity and number of functions. USCAR likewise indicated that ``the
most difficult issues lie not in the design of the software in
individual modules, but in the interactions between different modules
and components--i.e., integration of embedded systems composed of
heterogeneous components designed and implemented by different
suppliers.''
Cyber-physical systems are pervasive in other military systems.
Emerging systems (manned and unmanned) are incorporating greater
intelligence and autonomy. Collaborative, network-enabled operations
between multiple systems are becoming the rule rather than the
exception. The CBO (published in, ``The Army's Future Combat System
Program,'' April 2006) has indicated that at least 34M lines of
software code, much of it for CPS, will be generated for Future Combat
Systems--about twice current estimates for the Joint Strike Fighter.
Today's generation of fighters (figure below) incorporate many cyber-
physical systems. These systems operate in highly dynamic environments
with real-time mission specific behaviors. This imposes challenges on
the cyber-physical systems in the areas of networking, information
management, verification, validation, and certification, to mention a
few.
Future aerospace systems will require cyber-physical systems of
even greater complexity. Systems will operate with autonomy and will
collaborate among themselves to provide vast gains in operational
effectiveness. Enabling capabilities in active resource management,
dynamic scheduling, and software enabled control mode changes will be
needed to support these behaviors. Systems of this sort have flown
today in research focused demonstrations. They will be the norm in the
future.
Estimates on source lines of code for systems beyond the current
generation of developing systems are several orders of magnitude
higher--and will likely exceed one billion lines of code.
Requirements for cyber-physical systems and software are far more
stringent than those for typical office automation applications. Our
systems must support real-time behavior. We require ultra-high
reliability and many of our systems are safety critical and require
certification by the FAA or equivalent military authority. While the
occasional ``Blue Screen'' may be painful in the office environment, it
can have extreme consequences in the air. Many of our military systems
need to be designed to support coalition operations with multi-level
security requirements. Our systems must also be hardened to withstand
future cyber attacks by adversaries. Because of these unique
requirements and the relatively small numbers of systems, we do not
expect a large investment from the commercial IT sector in these
technologies.
In order to achieve these cross-cutting capabilities, we will need
advances in technologies such as model-based development tools,
methods, and validation environments to build systems rapidly and
affordably. Moreover, we will require product-focused technologies
including software reuse, architectures, real-time theory, languages,
and product line architectures to achieve system affordability by
recouping investment across multiple system developments.
We have achieved some measure of progress. Several years ago,
Boeing developed middleware-based product line architectures to support
our military system developments. Sizable investments were made in new
CPS architectures and infrastructures (e.g., Bold Stroke and the FCS
System of Systems Common Operating Environment) and substantial gains
in productivity were realized. The middleware-based approach is
critical since the days where military systems lead and dominate the IT
industry are long past. Specifically, CPS architectures like Bold
Stroke (illustrated below) were developed in part to provide layers of
isolation between the avionics software for DOD systems like F/A-18 and
F-15 from hardware and operating systems from the commercial IT
industry.
The challenges today are far greater than those faced in even the
recent past and continue to grow as individual systems evolve, operate
with greater autonomy and intelligence, and operate as part of a
networked system of system. The challenges grow even larger with future
generations of unmanned air systems operating in national air space.
What is the way ahead? Efforts to date have largely been fragmented
across the industry and limited by internal funding constraints. CPS
investments cross multiple technology domains and will require
industry-level critical mass to achieve the needed results.
We need a national strategy in which long-term CPS technology needs
are addressed by combined government and corporate investment. Boeing,
for its part, can focus our long-term CPS investments on collaborative
research in which we provide challenge problems and in-kind
participation in government-industry research consortiums. I'm
confident our industry partners are willing to do the same. We also
need to develop new ways to facilitate the transition of research
products back to industry and into our products. This point is critical
and is a matter of national competitiveness. The European Union's (EU)
Advanced Research and Technology for Embedded Intelligence and Systems
(ARTEMIS) program is funded by a public-private investment (over $7B in
mid-2007 dollars) and is pursuing R&D to achieve ``world leadership in
intelligent electronic systems'' by 2016. European industry is fully
partnered with academia in ARTEMIS. From our perspective, partnership
between industry and academia in CPS is absolutely essential to reap
the benefits of this advanced research. This partnership needs to reach
deeper than the rather ``indirect'' approach used for industry
involvement today.
In summary, we support the proposed expansion of the NITRD
program's research objectives to address cyber-physical systems and we
look forward to the opportunity to participate.
That concludes my testimony. I'd be pleased now to respond to your
questions.
Biography for Don C. Winter
Don has been employed at Boeing and its predecessor companies for
31 years. He holds BS and MS degrees in Physics from the University of
Missouri and an MBA from Washington University. Don held a number of
avionics design and systems engineering assignments on the Tomahawk
cruise missile program from 1977 to 1988. He joined the Mission
Planning Division of McDonnell Douglas Missile Systems Company in 1988,
serving in program management roles on the Tomahawk Mission Planning
Upgrade and (UK) Automated Mission Planning Aid (AMPA) programs, 1988-
91. In 1992 he was named Deputy Program Manager for the USAF Air Force
Mission Support System (AFMSS). In 1995, he joined the Production
Aircraft Advanced Design organization as Manager--Mission Systems, and
founded the Common OFP initiative, which later served as the foundation
for the Bold Stroke advanced avionics program. He led CRAD programs
under the Phantom Works Open Systems Architecture technology thrust
from 1998 to 2001, when he became Director of the overall thrust. He
then led the PW Network Centric Operations thrust, focused on the
development of key technologies and tools for network enabling Boeing
systems and products, from its inception in 2003 through mid-2005. Don
then assumed leadership of the Integrated Command and Control
organization within the C3ISR Solutions business segment of Boeing
Integrated Defense Systems. In this capacity Don led the development
and execution of C2 market shaping, product development and business
pursuit strategies. Most recently, Don returned to Phantom Works to
lead the Engineering and Information Technology organization, a team of
1,000 engineers and scientists performing leading edge research
spanning domains ranging from aeronautics and propulsion to avionics,
sensors and advanced information technology. Don has authored numerous
technical publications and currently serves on advisory boards at the
University of Cambridge, the University of California-Berkeley,
Vanderbilt University and Washington University.
Discussion
Chairman Gordon. Mr. Miller, as you know, the PCAST top
recommendation was the cyber-physical security, and we do want
to get more involved in that.
At this point we will open the first round of questions.
The Chair recognizes himself for five minutes. What I would
like to do is I have a couple of questions just to put to the
panel in general, and the first is the PCAST assessment of
NITRD program indicates that it should rebalance the funding
portfolios by increasing support for important problems that
require large-scale, longer-term multi-disciplinary R&D,
increasing emphasis on innovation and therefore high risk but
potentially higher payoff expirations. Do you agree with this
recommendation? If so, what should be done to implement support
for such large-scale innovative research projects? Dr. Greer,
why don't we start with you?
Dr. Greer. Thank you, Chairman. I think we would agree with
that recommendation of the PCAST assessment that these
investments in high risk but high payoff in multi-disciplinary
undertakings are important, critical in the current IT
landscape. As I said, our strategic plan focuses on identifying
challenges that can only be approached by agencies working
together and the corresponding technical issues to achieve
that.
Chairman Gordon. And can you do that on existing funding or
does that require additional funding?
Dr. Greer. I think we would look to both of those
possibilities, that refocusing some existing funding on these
shared projects, identifying those things that would
substantially enable an agency's mission if it could be
accomplished. Then that merits some focusing of funding. There
may be other opportunities for added funding in that same
category.
Chairman Gordon. Will you be identifying those areas of
opportunities and funding?
Dr. Greer. Yes, that is part of our strategic planning
process to look for what are the categories of this type, the
multi-agency, higher-level challenges that could be taken on by
a joint effort across agencies in which the missions of the
individual agencies are supported by this joint effort.
Chairman Gordon. Would anyone else like to comment on that?
Okay.
Dr. Reed. Well, I should say I hope I agree with the
commendation in this case since I helped write it. Just a
reminder that the PCAST was not the first time that this
observation has been made. If you go back and you look at the
1990--report, the last systemic evaluation of the program, it
recommended something quite similar, something that was called
out as expeditions to the--intended to be large-scale
integrated investments--technology that could have a
transformative effect. And I agree with Dr. Greer, it will be
quite a mix of targeted reallocation and most likely some
additional investment.
Chairman Gordon. Has this recommendation been made before
or are they behind the curve on getting something done?
Dr. Reed. There were responses to those recommendations,
perhaps not at the scale that the original recommended, some
due to some financial constraints. I think it is difficult to
change the culture of investment because again, it is not just
an agency response, it is a community response to the changes
that the Federal Government induces. And there are strong
incentives among the community to continue in many cases the
status quo. So it is not just a government issue, it is a
community education issue about the--and the risk of----
Chairman Gordon. I only have five minutes also, so let me
get to my second question. PCAST also recommended that NITRD
Program provide increased support for research on software but
does not cite specific research needs where the current program
is deficient. Software is a perennial area of weakness, and
information technology and an area in which NITRD Program
currently allocates resources. What is missing is software
research resources limited or idea limited? And what research
is needed to make greater progress for improving the
capabilities and reliable software. And I think, Dr. Reed, we
should probably start with you on that one.
Dr. Reed. I think it is a question of scale, and to hark
back to something that Mr. Winters said, as an example, cyber-
physical systems. What is happening is our software systems are
growing exponentially in size and complexity but perhaps even
more worrisome is that they are not isolated. They can control
our physical environment in all kinds of day-to-day ways, from
national infrastructure to our personal experiences. And the
search portfolio, back to the previous question, I think the
challenge is to look in at how we address software at large
scale. There is lots of research on the small-scale software
issues but how to deal with large, complex systems where a
small group is unlikely to understand its behavior, its
reliability, and its dynamics. There are some deep research
issues there. Some innovation is required. There are questions
frankly we don't know the answers to. It is not even in some
cases clear how to approach solutions to the problems.
So the basic research, but there is also a scale issue
about approaching a problem that is challenging at the moment
in academics.
Chairman Gordon. If I can ask, and I will try to be quick
with this, this is the situation we run into so often is there
is simply not enough money being invested. Now, we can get
into, you know, what are best parties, that sort of thing. I
think we need to do it through efficiencies in two ways.
Certainly the interagency is an excellent approach, and I
compliment what you have done. The second is whether or not
these are appropriate areas for international cooperation. You
know, what is our parochial interest here or first to market
interest versus international collaboration in terms of trying
to bring some economy to the research? Would you all give me
some quick thoughts on that?
Mr. Winter. I will give a couple of thoughts. I think that
we have examples already of international cooperation in our
business in areas considered to be infrastructure issues
versus, you know, areas of proprietary or competitive
advantage. We are not in the software business or the IT
business. Fundamentally, we are in the system business, the
aerospace system business. And we believe areas such as cyber-
physical system, infrastructure investment, is something we are
willing to do in a collaborative basis with our competitors,
with academia, and with international----
Chairman Gordon. Well, let me ask you a quick question
again. Is there any existing vehicle for that type of
collaboration now, any multinational agency or anything that
you know of? Dr. Greer.
Dr. Greer. This is an important question. There are areas
of NIT that are inherently international. The Internet itself
is global. Cyber security is a global issue. In each of those
areas, an area of large-scale networking, for example, there
are a series of organizations that manage the standards
development that manage the network operation and so on. So in
each of those separate areas, there are international
organizations. I think the challenge is in coordinating across
those international bodies, particularly in the area that Dr.
Reed has described, fundamental research, the software science,
the theory of mathematics.
Chairman Gordon. My time is up, but I would like for you if
you have an opinion to respond to us again in this aspect of
limited resources but not limited needs, where you think there
might be areas for international cooperation, you know, where
they are already going on, those other agencies, and should
there be, you know, whatever coordinated body. In other words,
how can we get better bang for our buck here without harming
ourselves in a first to market or proprietary way.
Thank you, and excuse me, Mr. Hall, for taking a little
time there. You are certainly recognized.
Mr. Hall. You are the Chairman. Mr. Winter, you mentioned
ARTEMIS. You mentioned the need for the United States to have
something similar to the EU public/private partnership,
research, technology, and so on and so forth called ARTEMIS.
Although termed an EU program, isn't it true that industry
contributes more than half of the funding for this and was
responsible for the total start-up and operational costs of
ARTEMIS? Is that true?
Mr. Winter. Yes, this is under the European Framework
Program which is a model for collaborative government, private-
sector investment.
Mr. Hall. Do you think the U.S. industry would commit to
the same level?
Mr. Winter. Yes, I do.
Mr. Hall. Would Boeing?
Mr. Winter. Yes.
Mr. Hall. Dr. Reed, would Microsoft?
Dr. Reed. Microsoft already is in many areas, so I have no
doubt about that.
Mr. Hall. You think our own industry is up to funding at a
greater extent?
Mr. Winter. I think we are already doing it to a large
degree and it is a matter of coordination, being provided by
nationally led activity, supplemented by some public funding
for the academic sector. I think it is more of a matter of
channeling investments we are already making.
Chairman Gordon. Mr. Hall, you raised a very good point.
Could you also respond to the Committee on how we could do
that, your suggestions in that area? I think Mr. Hall raises a
very good point. Thank you.
Mr. Hall. Do we await that? Oh, no, no.
Chairman Gordon. No response. They will get it to us in
writing.
Mr. Hall. All right. Okay. I think my time is up. I yield
back.
Chairman Gordon. Mr. McNerney is recognized for five
minutes.
Mr. McNerney. Thank you, Mr. Chairman. First of all, I want
to remark on a comment Dr. Stewart made. You recommended
consistent federal funding, and I just want to say that I feel
your pain on this. I spent my career in the renewable energy
business, and the production tax credits came and they went and
the industry suffered immensely from those cycles. And I am
sure that federal funding on a consistent basis would be better
than just about anything else. But then I also echo the
Chairman's words on this, we can authorize all we want in this
committee. If there is not enough money in the kitty, it is not
going to happen. And so Dr. Reed mentioned that 86 percent of
the funding is coming from one single federal agency, and that
has its advantages because it allows better coordination but it
has a disadvantage. I didn't quite understand what the
disadvantages were if you would elaborate on that a little bit,
Dr. Reed?
Dr. Reed. Certainly. Historically as I said, the diversity
of agencies had different approaches, and by nature of the
agencies and because they have differing missions and often
there is a pipeline of technology process and research that
leads to its impact in commercial industries. If you look, for
example, at some of the major technologies that we take for
granted now, there is about a 20-year pipeline from basic
research until they become billion dollar or more industries.
But in the federal context, that often meant a curiosity
driven, single investigator research cannot get by, researchers
are typically academia funded by the National Science
Foundation. DARPA on the other hand tended to focus on much
more goal-directed outcomes, larger scale projects, building
advanced prototypes in collaboration with industry. But it
built on ideas that had often been explored years before by
researchers funded by the National Science Foundation. So that
interplay of agencies meant that there were different ideas
that could be picked up and explored with different mechanisms.
It goes back to the interagency collaboration, coordination
mechanism, how that diversity created a variety of approaches
to innovation. And what has happened is we have lost some of
that diversity. It certainly in the academic side--on that as
opposed to broadly.
And so we only have the first order a single approach in
academic circles--that has been largely a single faculty
member.
Mr. McNerney. You were talking about sort of an agenda. If
it is a single agency, it tends to be agenda driven rather than
giving a diverse set of rules?
Dr. Reed. Right. You would like multiple goals, multiple
agendas, multiple kinds of approaches to select projects for
funding because that leads to different mixes of people,
different kinds of outcomes, and we have moved much more toward
a single kind of outcome model where it had been a much more
diverse model.
Mr. McNerney. It seems to me that if you have a well-run
program within the NSF, it would accomplish those objectives
rather than to try and coordinate different agencies.
Dr. Reed. It's a balancing act for sure, and I am sure Dr.
Greer can speak to that as well who balances that process. But
each agency has a different culture, and its culture makes it
easier or more difficult for it to do certain things. Some
things are much easier to do in a Defense Department model,
some things are much easier to do in a NSF-style model. And it
is that culture that the agencies struggle with when we try to
foster interagency collaboration and agency agendas versus the
broader sort of integrated agency and defense IT research.
Mr. McNerney. Okay. Thank you. Mr. Winter, I am interested
in your discussion of cyber-physical systems. Could you
elaborate a little bit on the current threat? What does that
look like, how immediate is it, do we have tools to move
forward aggressively on that?
Mr. Winter. You are referring specifically to cyber attack
on----
Mr. McNerney. Physical cyber attack, yes.
Mr. Winter. Cyber-physical systems are what we used to call
embedded systems, traditionally very isolated and stand-alone
entities. Apply control computer in an aircraft, for example,
wasn't subject to any external influence or attack. As these
systems evolve to a more collaborative model where they are not
only doing hard real-time business on board the platform but
are also participating as clients in a network, they become
vulnerable to cyber attack, tampering. The science for cyber
security for cyber-physical systems is really in its infancy.
Because the systems have only recently been sort of opened up
and made accessible because of their need to again service
clients on network, we are just in the early days of beginning
to really take on the cyber attack threat for these kinds of
systems.
Mr. McNerney. Well, I guess my time is expired, so I will
yield back.
Chairman Gordon. The gentleman's time has expired, and Ms.
Edwards is recognized.
Ms. Edwards. Thank you, Mr. Chairman. Just one question. I
mean, each of you raised the issue of technical capacity,
training, where are the professionals for the future, and going
to this question around cyber security, I read an article
yesterday about the use of virtualization, particularly in the
retail industry. And I wonder what the impact is, specifically
as you just mentioned, Mr. Winter, of the idea of
virtualization in these systems where you are trying to achieve
efficiency and, you know, monitor operations but how vulnerable
does that leave us to cyber attacks and what capacity do we
have to address emerging issues in technology? It just seems
like there is one every day around security. What is the
capacity we have right now to address those emerging issues
given the lack of capacity in the industry and in the federal
sector?
Mr. Winter. I think we do have a capacity issue.
Information assurance specialist is one of the most sought-
after and rare commodities from a field personnel standpoint in
our industry. We have a few, and we bend over backwards to keep
them with us. It is a small and slowly growing pool of
specialists, and I think the lack of collective training and a
qualified workforce in that area is a real threat to our
business. And to many other business, the financial sector,
other aspects of our national IT infrastructure.
Ms. Edwards. Dr. Reed, do you have a----
Dr. Reed. Oh, I have to agree with Mr. Winter. There is a
shortage of talent in this area, and as I said, because so many
of our everyday objects now include embedded intelligence and
network connections, there are substantial challenges here.
Microsoft, for example, has made trusted computing a major
initiative, to look at how to make software more robust because
as you have observed rightly, it is subject to a wide range of
attacks every day. Part of this is an inevitable consequence
that software permeates almost everything, but it is also an
absolute hard fact of the original question that Representative
Gordon asked about the challenges we face in building large
complex infrastructure and the underlying research issues
behind those. This is one example of the manifestation of that
struggle to build systems from first principles that are
reliable and secure, and the struggle is to retrofit security
systems as we discover vulnerability. But there are major
research issues here, workforce issues, and other venues people
have testified, and there are new programs to advance the state
of cyber security research under way now but there is a lot of
work to do without that.
Ms. Edwards. And so then to each of you, I mean, how do you
then prioritize where the allocation needs to go for research
and development, what areas, because it seems very expansive
and you know, it is clear from your earlier testimony,
obviously we haven't been able to fund everything and we won't.
Dr. Stewart. Funding everything is clearly beyond reason,
but I do think it is important to break up the funding
portfolio and look at it in terms of a variety of topics.
Cyber-physical systems, basic research and networking, research
networks in support of research in other areas cyber
infrastructure, production cyber infrastructure delivered
today, and the development of new cyber infrastructure for
tomorrow. And the threads of those research and development
activities need to go on continually so that the expertise that
we build up which is so precious gets retained in these
programs, and then while building the better workforce is, you
know, a 20-year process, from a 10-year-old to a 30-year-old
seasoned professional, looking at the long range and really
focusing on, okay, we recognize that today there shortages of
workforces. There are short-term measures that can be put into
place that will aid that in the short run, but really
fundamental efforts as Dr. Reed has already mentioned have to
be put into encouraging young people in the United States to
pursue careers in networking information technology.
Ms. Edwards. Thank you, Mr. Chairman.
Chairman Gordon. I thank you for your value added, Ms.
Edwards, you bring to our committee.
Mr. Hill, you are recognized for five minutes.
Mr. Hill. Thank you, Mr. Chairman, and I want to thank the
panel members for being here this morning, in particular Dr.
Stewart, who is not only chairman of CASC but is also the
Associate Dean at the greatest university in the history of the
world, Indiana University. Dr. Stewart, it is great to have you
hear today. Dr. Stewart, you mentioned in your testimony the
importance of consistency, and I want to return to that issue.
Are you suggesting that there are inconsistencies in funding
that are occurring?
Dr. Stewart. I think if you look back at the past several
years of funding in NITRD as implemented by the participating
agencies, there have been recognitions of areas where
additional funding is needed, and funding has been propped up
in one area at sometimes the expense of other areas; and that
oscillation in funding really creates difficulty in maintaining
the expertise base among people who really desire to pursue a
career in publicly funded networking information technology
research. If a program is eliminated or funding is temporarily
suspended and a person leaves publicly funded, publicly
oriented network information technology research, they are not
likely to come back. That expertise is lost, and expertise as
we have heard from I believe all four of us this morning is
tremendously valuable.
Mr. Hill. And this is happening?
Dr. Stewart. This has happened. Yes. Very definitely.
Mr. Hill. Okay. Switching gears then, one recommendation
you make is to increase the coordination between federal
agencies. How would you suggest that we do this or how would
you suggest this be done?
Dr. Stewart. Well, I think the key point is to add to the
coordination between federal agencies and add to that more
coordination colleges and universities, State investments and
regional investments, and the Coalition for Academic and
Scientific Computation and the Educause Campus
Cyberinfrastructure Working Group just held a workshop last
week in Indianapolis to generate new ideas specifically on this
topic. As Dr. Reed said, a lot of these issues have to do with
the culture of academia, and I think one of the key points is
to make recommendations both to academia and State-funded R&D
activities that enable them to better collaborate with
federally led initiatives and to add to the federally led
initiative ways that will allow that collaboration to make it
more effective.
Mr. Hill. Well, could the NITRD planning in coordination
mechanism be used as a forum for this broader level of
coordination?
Dr. Stewart. I do indeed. I think that is actually the best
way to begin that.
Mr. Hill. Thank you, Mr. Chairman, I yield back.
Chairman Gordon. Well, before we thank our witnesses, I
just thought I would check and see if anyone wants to have an
alternative opinion as to the world's greatest ever colossal
university. Dr. Baird.
Mr. Baird. I would just point out I am proud to have two
representatives from the great State of Washington here and let
that speak for itself.
Chairman Gordon. Again, I want to thank our witnesses. We
are trying to build a base of information. You have given us a
good place to move forward. We would hope that you would
respond to those questions that we ask as well as anything
else. As we go through this process, you know, there are a few
of us here but there are hundreds of thousands, if not--we
normally have a couple of million that watch our webcast, so
there are lots of others that are out there, and we want to
welcome any suggestions to this very important concern. And
this hearing is adjourned.
[Whereupon, at 10:56 a.m., the Committee was adjourned.]
Appendix:
----------
Answers to Post-Hearing Questions
Answers to Post-Hearing Questions
Responses by Christopher L. Greer, Director, National Coordination
Office for Networking and Information Technology Research and
Development (NCO/NITRD)
Questions submitted by Chairman Bart Gordon
Q1. Are there areas of IT research that are good candidates for
international cooperative efforts that would leverage U.S. investments
but would not otherwise harm U.S. competitive advantages, such as being
first to market for a new technology? What mechanisms are available, or
could be instituted, to facilitate such international cooperative
research?
A1. Because we live in a global digital society there are international
implications across all areas of networking and information
technologies (NIT). However, international cooperation is exceptionally
important in four NIT areas.
1) Long-term data preservation and access/sharing
This is a vital area for international cooperation, both because
21st century science is global and data-driven, and because many of our
society's challenges (e.g., energy and other natural resources, climate
change, biodiversity, and human health in a globally-connected world)
require coordinated international data sharing and analysis.
Current NITRD examples:
a) DOE/SC and NSF support international high-energy
physics research, including analysis of gigabytes-per-
second data from the Large Hadron Collider (LHC)
located at the CERN site near Geneva. This includes
high-capacity optical network links to two tiers of
U.S. analysis sites and development by DOE/SC of new
and extended protocols enabling massive data
throughputs across broadband network links.
b) The NSF-supported Very Long Baseline Array (VLBA)
of telescopes cooperates with the European VLB Network
to create a global Internet telescope with
unprecedented resolution for distributed near-real-time
observation and data gathering.
Possible NITRD cooperative partner on data issues:
The mission of the International Council for Science
(ICSU) Committee on Data for Science and Technology
(CODATA) is ``to promote, throughout the world, the
evaluation, compilation and dissemination of data for
science and technology and to foster international
collaboration in this field.'' The U.S. National
Committee for CODATA links the scientific and technical
community in the United States and the international
CODATA on data issues and operates within the National
Research Council's Board on International Scientific
Organizations. I have met several times with USNC
CODATA to share plans and information and look to
strengthening that link as a conduit for international
partnerships in the data arena.
2) Advanced networking
International networking cooperation is a prerequisite for seamless
global high-speed communications, including scientific data sharing.
The NITRD agencies thus have longstanding cooperative relationships
with international networking organizations and the scientific networks
of other countries, and continue to expand these partnerships.
Current NITRD examples:
a) The September 28-30, 2008 Networking Research
Challenges Workshop (with international participants
from the Netherlands, Canada, Japan, South Korea, and
China), will be held in conjunction with a meeting of
the Global Lambda Infrastructure Federation (GLIF),
which coordinates international cooperation and
transparency among the world's optical networks. One of
the key goals of this workshop is to explore the
international implications of the recently-developed
Federal Plan for Advanced Networking (see
www.nitrd.gov/ITFAN-preprint-061108.pdf).
b) NSF's International Research Network Connections
(IRNC) program, includes
a) The TransPacific Network (TransPac)
providing connectivity among Asia, Europe, and
the U.S.;
b) PerfSonar, a global collaboration among
national research and education networks in the
U.S., Europe, Latin America, and Asia that is
developing a distributed network measurement
framework to improve end-to-end performance for
researchers; and
c) the Pakistan-U.S. Research & Education
(R&E) Network connection, online as of August
15, which connects Pakistan's scientific
research and education community for the first
time to the U.S. and global R&E networking
fabric (also supported by the European Union's
TEIN2 project).
c) NSF, through its Network Science and Engineering
program, has been working with the U.S. academic and
industrial communities to create opportunities for
possible federation of experimental network
infrastructure, participating in the National Institute
of Information and Communications Technology (NICT)
JGN2 and AKARI (Japan's new generation network testbed)
Symposium in Tokyo in January 2008 and the Future of
the Internet Conference in Slovenia in March 2008. NSF
staff will participate in the launch of the EC's Future
Internet Research and Education (FIRE) projects in
Paris in September 2008. A joint Japan-U.S. workshop is
planned for October 2008.
d) LSN's Joint Engineering Team (JET) provides a forum
for the development of operational policies and
practices, including security policies and responding
to security incidents, on the international science
networks.
e) New techniques for inter-domain signaling developed
under the NSF-supported DRAGON project are enabling a
new networking paradigm--hybrid networking--that
combines shared IP services with dedicated high-
capacity capabilities for data-intensive scientific
research. The U.S. academic community's Internet2
consortium is deploying the Dynamic Circuit Networking
(DCN) service and ensuring international inter-
operability through active collaboration with a peer
network in Europe called GEANT.
f) To harmonize deployment of Internet Protocol
version 6 (IPv6) in Federal networks with existing
international programs, specifically the IPv6 Ready
Logo program, NIST has signed Memoranda of
Understanding for cooperative development of test
materials with members of the IPv6 Forum. Forum members
include: Yokogawa Electric Corporation of Japan, NTT
Corporation of Japan, NTT Advanced Technology
Corporation of Japan, Yaskawa Information Systems
Corporation of Japan, Institut National de Recherche en
Informatique et en Automatique (INRIA) and the
University of Rennes in France, as well as the Inter-
operability Laboratory of New Hampshire.
These examples illustrate how individual NITRD networking
activities are directly linked to the appropriate international
counterparts. Through its strategic planning activities, the NITRD
program will explore whether other organizations, such as the
Organization for Economic Cooperation and Development (OECD) through
its information and communication technologies activities, can provide
mechanisms for addressing the broader international networking issues.
3) Software engineering
Software is pervasive in our digital world, underlying the
operation of planes, ships, factories, and medical devices; and
controlling critical infrastructure such as power grids and banking and
financial systems. As these critical software applications become
increasingly complex, the need for robust software science--theory,
concepts, and methodology for creating, analyzing, and verifying
software--has become a global challenge.
Current NITRD example:
The Verified Software Initiative (VSI), a long-term
cooperative, international project directed at the
scientific challenges of large-scale software
verification. The VSI resulted from the first Verified
Software: Theories, Tools, Experiments Conference
(VSTTE) held in 2005 in Zurich as a response to Sir
Tony Hoare's (Microsoft Research U.K.) Grand Challenge
on the ``verifying compiler,'' a vision of software
produced with machine-verified guarantees of adherence
to specified behavior. International working groups are
in place and much work has been done, resulting in
multiple technologies now available to address the
challenge. The second VSTTE conference will be held in
Toronto in October 2008; agencies in NITRD's High
Confidence Software and Systems (HCSS) coordinating
group--including NASA, ONR, NSA, and NSF--are co-
sponsors of the VSTTE conferences and/or the grand
challenge technical activities, such as the
Verification Grand Challenge. These activities also
draw participants from Australia, Belgium, China,
Denmark, France, Germany, India, Israel, Japan,
Switzerland, and the United Kingdom.
This example illustrates how international cooperation can
accelerate progress on some of the most difficult, and most pressing,
NIT challenges of our time.
4) Embedded and cyber-physical systems
A modern automobile is an integrated cyber and physical system,
relying on embedded IT to control engine functions, anti-lock braking,
transmission, emissions reduction, vehicle stability, and entertainment
and climate control systems. Embedded NIT systems in planes, trains,
ships, traffic control systems, and emergency response networks are an
essential part of our everyday experience. Ensuring that these cyber-
physical systems, are safe, effective, predictable, and reliable is
another key global challenge.
Current NITRD example:
The High Confidence Software and Systems (HCSS)
Coordinating Group within the NITRD Program has a
strong focus on cyber-physical systems. Recent
workshops sponsored by HCSS agencies in this area
include high-confidence medical devices and systems and
automotive safety. The HCSS group is also active in the
international arena, participating in events such as
the Cyber-Physical Systems (CPS) Week at the annual
multi-conference (Real-Time and Embedded Technology and
Applications Symposium [RTAS], International Conference
on Information Processing in Sensor Networking [IPSN],
and International Conference on Hybrid Systems [HSCC])
jointly sponsored by IEEE and the Association for
Computing Machinery (ACM) and the upcoming Embedded
Systems Week (International Conference on Embedded
[EMSOFT], International Conference on Compilers,
Architecture, and Synthesis for Embedded Systems
[CASES], and International Conference on Hardware/
Software Co-design and System Synthesis [CODES+ISSS])
jointly sponsored by the IEEE, ACM, and the Council on
Electronic Design Automation. In the past month, a
jointly-sponsored U.S./Japan workshop focused on human-
robot interactions, emergency robotics, and medical
robotics generated considerable enthusiasm for the
potential for collaboration among investigators.
In summary, the HCSS group is developing both a national and
international coordination effort in the area of cyber-physical
systems.
Q2. Do you have suggestions for how the U.S. could institute a public-
private research partnership to advance the capabilities of cyber-
physical systems analogous to the EU's ARTEMIS initiative? Could such
an undertaking be planned and carried out under the NITRD program?
A2. Certain elements of the ARTEMIS model are uniquely shaped by its
European Union context. A model launched in the U.S. is SEMATECH
(SEmiconductor MAnufacturing TECHnology). This broad industry
consortium, which celebrated its 20th anniversary in 2007, began with
close partnerships with the Sandia and Oak Ridge national laboratories.
It is credited with restoring U.S. leadership in the industry. Thus,
public-private partnerships can have a deep and lasting impact on the
NIT landscape. The following examples, including one from outside the
NITRD Program, may suggest some possible starting points for new,
partnered initiatives.
Current NITRD examples:
a) In addition to federal agency networking
organizations, the NITRD/LSN's Joint Engineering Team
includes Internet2 and National LambdaRail (NLR) (both
consortia of university network organizations), along
with commercial entities such as Cisco, Juniper, and
Sun.
b) LSN's Middleware And Grid Infrastructure
Coordination (MAGIC) Team includes representation among
public and private science networking organizations to
provide inter-operability among grid infrastructures
(Open Science Grid, Open Grid Forum), regional and
local grid organizations, university grid capabilities
through Educause, and commercial-sector participants
such as Microsoft, IBM, and HP.
c) NSF's Industry & University Cooperative Research
Program (I/UCRC) fosters partnerships between academic
institutions, government agencies, national
laboratories, and industry, including IT-related
research centers reported in the NITRD crosscut. (For a
complete list, see (http://www.nsf.gov/eng/iip/iucrc/).
NSF's Cluster Exploratory (CluE) initiative has
launched two industry-academia-government
partnerships--one involving IBM and Google, the other
HP, Intel, and Yahoo!--to provide researchers access to
cluster computing resources.
d) Since 2005, NITRD's HCSS agencies have been
sponsoring a national workshop series on key domains
for cyber-physical systems (e.g., medical systems and
devices, aerospace and transportation, critical-
infrastructure and industrial-process control systems).
This series is expressly designed to bring together
public- and private-sector domain experts, researchers,
developers, vendors, and users to share their
perspectives and forge a common understanding on
research and user needs in these vitally important
technologies. These emerging communities of interest
could possibly serve as a basis for a more formal
partnership activity.
Related non-NITRD example:
In 2005, the Department of Energy, in collaboration
with the Department of Homeland Security and Natural
Resources Canada, partnered with industry leaders in
the electric, oil, and natural gas sectors to design a
unified framework to guide control-system cyber
security R&D efforts and investment. The resulting 10-
year Roadmap to Secure Control Systems in the Energy
Sector was published in January 2006. The newly formed
Energy Sector Control Systems Working Group composed of
public and private energy-sector leaders, has been
active in Roadmap implementation.
The SEMATECH example illustrates how an initial public-private
partnership focused on a critical NIT challenge can produce a
successful outcome. One of the largest technical challenges in
computing currently is how to design software that can effectively use
multicore systems. The PCAST cited the emergence of multicore
processors in its rationale for recommending that: ``the NITRD
Subcommittee should facilitate efforts by leaders from academia,
industry, and government to identify the critical issues in software
design and development and help guide planning on software R&D.''
NITRD agencies report the following recent initiatives in this
area: Intel has partnered with academia to open two new scalable
software research centers; DOD and DOE/SC have partnered with Goodyear
and Caterpillar to develop a new set of geometry and messing tools--the
foundation components needed to support fast problem solving on
multiple cores; and the Council on Competitiveness is working with
industry and NITRD agencies to organize a consortium for development of
scalable engineering applications. Scalable software may be an area in
which a significant level of shared public-private investment could
spearhead technical innovation that would benefit the U.S. economy as a
whole.
Q3. The COMPETES Act requires the NITRD program to develop and
maintain a research, development and deployment roadmap for high-end
computing systems. What steps are being taken and what is the timing
for implementing this requirement?
A3. The NITRD Program is currently in the fourth year of the five-year
plan set forth in the May 2004 Federal Plan for High-End Computing
developed by an interagency task force at the request of the Office of
Science and Technology Policy. Accomplishments to date achieved by the
NITRD agencies as called for in that Plan include: development of
leadership-class capability HEC systems; making cycles on those systems
available to the broader private-sector research community for cutting-
edge computational R&D projects; revitalization of R&D in HEC systems
software through the HEC-University Research Activity HEC-URA); and
collaboration on methods to streamline federal procurements and develop
system bench-marking and evaluation tools that specifically address
federal requirements.
Under the NITRD strategic planning process, all of the Program's
research areas including HEC will develop new research plans and
technical roadmaps coordinated with the overall vision laid out in the
NITRD Strategic Plan. The timeline for these activities is presented in
Appendix 3 of my written testimony.
Q4. One of PCAST's recommendations is for the NITRD National
Coordination Office (NCO) to be more proactive in communicating with
outside groups. Was this a fair criticism of the NCO and do you intend
to make any changes related to the recommendation?
A4. We do not view the PCAST's recommendation as a criticism. It
acknowledges the reality that the NCO must keep evolving in tandem with
the evolution and maturation of the NITRD enterprise as a whole; PCAST
has provided an opportunity for us to accelerate our efforts in a
direction we were already headed. I've mentioned NCO-supported outreach
efforts through the LSN teams and the HCSS workshop series. Another
strong example is the monthly Expedition Workshop series co-sponsored
by NITRD's SEW Coordinating Group. These workshops supported by NCO
draw upwards of 100 participants spanning government, academia, and
industry in an ongoing dialogue about ways to harness emerging
technologies to improve public and private services for citizens. It is
notable, for example, that the workshops have spawned more than a dozen
professional Communities of Practice across the Federal Government,
several of which have developed data standards adopted by the Office of
Management and Budget under the Federal Enterprise Architecture. An
automated emergency alert technology incubated in the workshops won
OMB's Federal Innovation of the Year award.
The R&D coordination responsibility assigned to NITRD under the
Comprehensive National Cybersecurity Initiative (CNCI) has at its core
outreach to and close partnerships with the private sector. To
accelerate the advance of new cyber security technologies toward
commercial implementation, as called for under CNCI, the NCO must work
aggressively to help agencies forge innovative working relationships
with private-sector researchers, developers, vendors, and technology
users. We have already taken key outreach steps, including having NCO
staff schedule meetings with industry officials during office travel,
and contacting industry representatives to participate in forthcoming
high-level brainstorming sessions with federal cyber security managers.
It is my expectation that the NITRD Program's strategic planning
process itself will identify additional opportunities for private-
sector outreach. Several outreach efforts--a request for public inputs,
a web site for public discussion of these inputs, and a national
workshop--are already incorporated into our activities to develop the
Strategic Plan for NITRD. In addition, I anticipate that the parallel
strategic plan we will develop for the NCO, as called for by the PCAST,
will specifically address new forms of NCO outreach activity in support
of the NITRD Strategic Plan.
Q5. Dr. Stewart described examples of collaborative distance education
to provide computer science related courses to students at institutions
that may not have strong programs in this field as one way to attract
more students to information technology careers. He suggests increased
support for such programs could be particularly valuable in increasing
the number of information technology professionals from under-
represented groups. To what extent are such collaborative distance
education programs now supported under the NITRD program, and are there
any impediments to increased funding for such activities?
A5. While broad implementation of education delivery systems lies
outside the NITRD Program's core mission, the R&D activities under the
program have resulted in many of the technologies and resources that
enable distance-education, supported research in best practices and
approaches, and examined the social and behavioral implications of
remote learning and virtual interaction. The NITRD Program and its
predecessors have been the vanguard of the technological revolution
that made distance learning possible and that continues to enrich its
capabilities. The NITRD agencies including DARPA, NASA, NIH/NLM, and
NSF (as well as the Library of Congress and the National Endowment for
the Humanities) played a lead role in developing the enabling
technologies for the concept of digital libraries and supporting the
creation of the first generation of major digital collections of human
knowledge and artifacts. Today, we take for granted that written works,
art, and historical artifacts are accessible to us online; in 1994,
when the NITRD agencies initiated their digital libraries activities,
such access was a dream.
Examples of technologies applied in distance learning that
originated in NITRD research include: modeling and simulation of
experimental data; haptic devices for remote manipulation of
instruments and visualizations; the Visible Human series of images of
the human body; multi-modal computer interfaces and interactive
devices; hyperwall technologies; grid technologies, applications, and
services; and wireless, hybrid, and all-optical networking
technologies, to name only a few.
The National Science Foundation--with its broad mission of support
for STEM education as well as for academic science and engineering
research--sponsors an array of formal distance-learning activities,
including projects reported under the NITRD crosscut. Under a recent
NSF award to the University of Houston, Downtown, The American Indian
Higher Education Consortium (AIHEC), the Hispanic Association of
Colleges and Universities (HACU), and the National Association for
Equal Opportunity in Higher Education (NAFEO) propose to establish the
Minority Serving Institutions (MSI)-Cyberinfrastructure (CI)
Empowerment Coalition (MSI-CIEC) to foster a CI-enabled distributed
education and research network providing e-science education and
research opportunities to MSI faculty and students. MSI-CIEC will
provide the ``human middleware''--the social and technological
mechanisms facilitating the necessary communication and support
linkages between MSI faculty and students, and researchers associated
with e-science and CI initiatives.
NIH supports professional distance learning through the National
Library of Medicine's (NLM's) PubMed and Medline digital archives as
well as through a growing assortment of biomedical image and data
collections and networks for sharing such information. The NLM also
supports experiments and training in telemedicine applications. The NIH
Office of Science Education provides online resources for educators, an
e-mentoring program for high school and college students, and career-
planning materials.
NSA and NSF are partnering with other funders in a North Carolina
program with distance-learning components that are applied to help
support exceptional K-12 teachers in STEM improve their curricula by
working with higher-education faculty and researchers. Through the
Kenan Fellows Program (KFP) for Curriculum and Leadership Development,
the competitively selected teachers spend two years conducting
experiments with researchers and developing new curriculum ideas and
techniques based on their work.
Non-NITRD examples:
Though not part of the NITRD crosscut, NASA operates what is
perhaps the Federal Government's most vibrant distance-learning
activity, in that the agency has incorporated outreach by means of
advanced digital technologies into its real-time explorations of Earth
and space, including vast archives of scientific images and broadcasts
from space. Microsoft's WorldWide Telescope and Google Sky are
innovative tools for experiential learning enabled by the data
resources of the Sloan Digital Sky Survey, the Hubble Space telescope,
the Wilkinson Microwave Anisotropy Probe, and the IRAS (infrared),
Chandra (x-ray) and GALEX (ultraviolet) missions.
Although it is not part of the NITRD crosscut, the Information
Resources Management College of DOD's National Defense University
offers 50 graduate-level courses that can be taken in a ``distributed
learning'' format. These courses focus on various aspects of
information technology leadership leading to certificates for CIO,
including: Information Assurance, IT Project Management, Organizational
Transformation, and Enterprise Architecture. These courses are
available to Federal, State, and local government employees who are
college graduates, and to government contractors. This provides an
excellent example of how the NITRD agencies are using distance learning
capabilities to meet their own education and training needs.
As noted in my written testimony, we are incorporating education
issues in the NITRD strategic planning process, and have initiated a
fast-track study of NIT education as recommended by PCAST. In addition,
NITRD's SEW Coordinating Group will hold a workshop September 16, 2008
to bring together representatives of Federal agencies, including non-
NITRD agencies, with responsibilities in the education arena. The
workshop will focus on both the role of NIT in education and NIT
workforce needs for the future. This meeting is intended to kick off an
ongoing education activity under NITRD's strategic planning process.
Answers to Post-Hearing Questions
Responses by Daniel A. Reed, Director, Scalable and Multicore
Computing, Microsoft Corporation
Questions submitted by Chairman Bart Gordon
Q1. Are there areas of IT research that are good candidates for
international cooperative efforts that would leverage U.S. investments
but would not otherwise harm U.S. competitive advantages, such as being
first to market for a new technology? What mechanisms are available, or
could be instituted, to facilitate such international cooperative
research?
A1. This is an extraordinarily complex problem, given the global nature
of information technology and the role that U.S. multinational
computing companies play around the world. As with all technologies,
one must chose carefully, leveraging the intellectual value of
international collaboration, while avoiding the loss of competitive
advantage in the U.S.
In an increasingly competitive environment, it is unlikely,
however, that the U.S. will maintain intellectual leadership in all
areas of computing. Thus, in certain areas, it is in our interests to
collaborate. For example, the European Union is now highly competitive
with the United States on formal verification and embedded systems
research.
In addition to specific technology areas, one might focus on the
applications of computing to international problems, climate change,
the environment and global health and nutrition. In these domains,
there are many computing research challenges, including data management
and mining, software and computer architecture. This is just one area
where international collaboration might benefit the U.S.
Regardless of the chosen areas, one must remember that the benefits
from research accrue disproportionately to the Nation in which that
research is performed. Because information, such as advancements in
basic science, is most easily communicated through interpersonal
interactions, having those interactions occur within our borders makes
it much more likely that U.S. industry will capitalize on those
advancements.
Q2. Do you have suggestions for how the U.S. could institute a public-
private research partnership to advance the capabilities of cyber-
physical systems analogous to the EU's ARTEMIS initiative? Could such
an undertaking be planned and carried out under the NITRD program?
A2. Yes, such an undertaking is possible. ARTEMIS is structured around
the European Union R&D processes, with much tighter academic and
industry collaboration than is typical in the U.S. To be globally
competitive, I believe the U.S. must reconsider some of the ways
industry-academic collaborations are currently structured and reassess
the reward metrics and associated intellectual property mechanisms.
Moreover, industrial-academic partnerships in U.S. have often been
difficult given the financial focus on quarterly returns. We must
restructure the compact among the parties in a way that industry
invests in an appropriate share of long-term research. Microsoft, for
example, is investing aggressively in long-term research, both via
Microsoft Research and via external funding to academic partners.
Cyber-physical systems span a broad spectrum, from national and
international infrastructure to home heating and cooling controls, and
would be a candidate for this type of partnership. This is an area
where opportunities might be best identified by a series of focused,
government-academe-industry workshops. The NITRD program would be best
served by picking one or two target areas that could be defined clearly
and focused on removing programmatic impediments to success.
Q3. Are the R&D objectives of the NITRD program, as it is currently
constituted prioritized appropriated and is the allocation of funding
consistent with achieving the objectives? Are there particular research
areas that the NITRD program is not pursuing with sufficient resources?
A3. No, not at present. We must increase investment in software and
cyber-physical systems (see below), even if it means decreasing funding
in some other areas.
In addition to software, we need better tools for managing the
explosive growth of computer-generated data. The era of the personal
petabytes is very near, and our mechanisms for ensuring long-term data
preservation, security and privacy are ill-suited to today's data
volumes, much less those expected in the future. Moreover, extracting
insight from such large volumes of distributed data remains
extraordinary difficult.
Finally, and perhaps most importantly, we must take a systemic,
scalable approach to integrated computing challenges. Many, arguably
most, of the computing R&D challenges require multi-disciplinary teams
of computing researchers--computer architects, hardware designers,
system software researchers, network visionaries, programming model and
tool experts, data mining and management researchers, and domain
experts.
Our current research ecosystem makes it difficult to both assemble
such teams and to sustain them long enough to mount explorations of
systemic challenges. It was for this reason that the 1999 PITAC report
recommended funding large-scale, revolutionary explorations--
Expeditions to the 21st Century. Such explorations, involving academe,
government and industry, are a missing element of our research
ecosystem and would go far to rebalance the risk-reward portfolio of
the NITRD program in favor of more long-term, high-risk research.
Q4. The President's Council of Advisors on Science and Technology
(PCAST) recommends that the NITRD program provide increased support for
research on software but does not cite specific research needs where
the current program is deficient. Do you have recommendations for how
the NITRD's investment in software research could be strengthened,
assuming no substantial increases in overall funding?
A4. There are at least four major software challenges before us today;
each is equally important.
I. Reliability and correctness of large software systems.
Much of our critical national infrastructure and our daily
lives depend on software systems--our financial markets,
communication systems, electrical power grid, transportation
infrastructure, signals intelligence, commercial web services
and enterprise software. Our lives and even our identities are
dependent on software systems that manage information and
infrastructure on our behalf, yet we do not have good methods
to ensure the reliability of these systems or to design them to
operate correctly--on time and on budget.
II. Cyber-physical system software models and tools. In some
sense, cyber-physical systems are a special case of the first
challenge, albeit with greater coupling of sensors, actuators
and communication via wired and wireless networks. From an
implanted pacemaker to the electronic fuel injection and anti-
skid brakes in today's automobiles though avionics in a
military or commercial jet, cyber-physical systems touch us
minute-by-minute. These distributed systems are ubiquitous,
because their advantages are manifold, but also difficult to
design and validate, given their complexity and the
interdependence among disparate components. Moreover, failure
of one system component can have far-reaching and often
disastrous consequences. (Consider, for example, the global
effects of a single design failure in a commercial jet's
avionics system.)
III. Security, privacy and resilience in an uncertain world.
Almost all of our critical data--personal, corporate and
government--reside in distributed data systems and networks.
Many of these systems are vulnerable to cyber attacks and to
malicious behavior. We must develop tools and techniques to
design more resilient systems, and ones that are provably
secure.
IV. Efficient and easy-to-use tools for multicore
programming. For over thirty years, we have been the
beneficiaries of a virtuous cycle of new and richer computing
systems, powered by ever faster microprocessors. Each new
generation of processors executed old software faster and
enabled new capabilities--graphical interfaces, speech
recognition systems and mobile devices.
Today, device power limits are forcing a new approach to
chip design--placing multiple processors on each chip. Such
multicore systems pose daunting challenges for software
development, requiring parallel programming to deliver high
performance on new applications. However, we lack the necessary
tools that would enable software developers to exploit these
multicore processors easily and effectively. This multicore
programming crisis is one of the deepest facing the commercial
software industry today, and inadequate research investment in
years past is one of our current problems.
Unless we find solutions to these problems--and soon--not only will
we risk catastrophic failure of critical national infrastructure, the
virtuous cycle of hardware-software innovation that has driven the
computing industry will be threatened. Some of these, such as ensuring
correctness and reliability in large, complex software systems, are
longstanding. Others, such as multicore programming, are more recent.
In a fixed budget scenario, we must reallocate funds from other,
lower priority R&D activities and better manage and coordinate extant
investments to increase efficiencies by eliminating redundant
activities.
Q5. Does the research community--both academe and industry--have a
voice in influencing the research priorities under the NITRD program?
Are improvements needed in the external advisory process for the NITRD
program?
A5. Yes, but not a fully effective voice. Academe influences research
priorities via workshops and joint meetings with the NITRD program
agencies, but the relative investments in specific technical agendas
are more often driven by agency needs than by community priorities. The
NSF, as a research agency, is perhaps the most responsive to community
priorities but even there proposals for specific budget allocations are
rarely discussed with the community.
On the industry side, there are fewer mechanisms for community
engagement with NITRD agencies. In general, industry trade associations
tend to represent short-term issues, rather than the basic research
topics central to the NITRD portfolio. Individual companies do engage,
but do so carefully lest they be viewed as advancing parochial
interests.
I believe we must have greater coordination across the government-
academic-industry partnership. Our international competitors recognize
the critical importance of such partnerships. We cannot afford to
continue to treat the three partners as arms-length collaborators.
Hence, we need much more than pro forma workshops and meetings which
``rubber stamp'' extant agendas if we are to maintain our competitive
position. We also need to find new ways for collaborative technical
partnerships across government, academe and industry, including honest
assessments of intellectual property issues.
PCAST recommended that the NITRD program increase its strategic
planning and define roadmaps for realizing the strategic plans. These
planning exercises, together with public assessments of progress
against the plans, would be an ideal mechanism to engage academe and
industry to regularly scheduled assessments and recommendations.
Answers to Post-Hearing Questions
Responses by Craig A. Stewart, Chair, Coalition for Academic Scientific
Computing; Associate Dean, Research Technologies, Indiana
University
Note: these responses are presented by Dr. Stewart on behalf of
CASC. These responses have been endorsed by a majority vote of CASC
members, without dissensions.
Questions submitted by Chairman Bart Gordon
Q1. Are there areas of IT research that are good candidates for
international cooperative efforts that would leverage U.S. investments
but would not otherwise harm U.S. competitive advantages, such as being
first to market for a new technology? What mechanisms are available, or
could be instituted, to facilitate such international cooperative
research?
A1. As a general approach to networking and information technology,
CASC recommends focusing on international collaboration efforts that
will set standards for inter-operability. Examples include standards
set by the Open Grid Forum (http://www.ogf.org/) for grid computing and
international standards for networking. Over the next several years,
two areas should be priorities:
Advanced optical networking standards and techniques
for high-bandwidth connections (e.g., greater than 40 gigabits
per second), techniques for dynamic creation of dedicated
networks in support of virtual organizations, and development
of sensor networks for environmental and resource monitoring.
Identity management--particularly creation of
national online identity management systems for research
cyberinfrastructures, and the creation of international trust
relationships between such systems where appropriate. This
would greatly facilitate international collaboration across
many areas of science and technology. One example would be
establishing InCommon (http://www.incommonfederation.org/) as
the definitive U.S. credential management system for research
IT (as CASC has previously recommended).
Such efforts might be funded via international collaboration of
U.S. federal funding agencies and the European Union Research Framework
Programme (http://cordis.europa.eu/fp7/home-en.html). A
joint call with issues by the U.S. and the EU Research Framework
Programme, and joint funding, would be a highly effective way to
promote international collaboration in networking and information
technology.
The creation of internationally accepted standards for inter-
operability of networking and information technology systems is
essential to the U.S. and to its ability to cooperate internationally.
We believe that the U.S. then competes by developing the best
implementations of information technology, the most effective cyber-
physical systems, and the most effective interaction of simulation via
computer and verification through experimentation. In this way the U.S.
can collaborate when appropriate, and compete (and win) when competing
is appropriate.
Q2. Do you have suggestions for how the U.S. could institute a public-
private research partnership to advance the capabilities of cyber-
physical systems analogous to the EU's ARTEMIS initiative? Could such
an initiative be planned and carried out under the NITRD program?
A2. The ARTEMIS initiative is an excellent model, and establishing a
partnership based precisely on this model could and should be carried
out under the NITRD program. A critical element for success of such a
program in the U.S. would be to fund time for U.S. academic researchers
to work as part of such collaborative efforts. The most valuable
commodity brought to private-public partnerships is the time of the
public sector experts and researchers. U.S. university and college
technology transfer offices tend to focus on intellectual property
outcomes resulting from collaborative research before such research is
even initiated. These negotiations are a significant obstacle to
public-private collaboration. This situation arises in part as a result
of the interpretations of the Bayh-Dole Act. For a public-private
partnership modeled on ARTEMIS to be most effective, it might be
helpful to include specific terms in a solicitation (with, if needed,
accompanying legislation) that facilitates partnership and innovation
by establishing clear guidelines for technology transfer to the private
sector and rights to and payment for same.
Q3. Are the R&D objectives of the NITRD program, as it is currently
constituted, prioritized appropriately and is the allocation of funding
consistent with achieving the objectives? Are their particular research
areas that the NITRD program is not pursuing with sufficient resources?
A3. The prioritization recommended in the PCAST report is, overall,
appropriate. (I note that in the written testimony CASC added increased
focus on complexity-hiding interfaces such as Science Gateways as a new
area for emphasis that has come to the fore since the PCAST report).
However, the objectives set forth in the PCAST recommendations
cannot be carried out effectively, in ways that preserve U.S.
international competitiveness, without substantial increases in the
NITRD budget.
CASC recommends a significant increase in the NITRD budget, with
focus particularly on three areas:
Creation of a national research cyberinfrastructure
as an interagency activity. Such a national cyberinfrastructure
should have significantly greater capability than the aggregate
of the various federal agency initiatives. We echo and support
particularly Dr. Reed's testimony on the point of interagency
funding balance. Because there are multiple large-scale
cyberinfrastructure efforts, and because none of the
cyberinfrastructure systems are yet straightforward enough for
most researchers to use, the number of researchers currently
using such advanced facilities is in the low tens of thousands.
U.S. global competitiveness would be best supported if hundreds
of thousands of researchers could use these facilities.
Rebalancing the NITRD budget so that a relatively
greater fraction of overall funding is devoted to software
development. Three areas stand out in particular: development
of parallel programming tools and applications for multicore
processors; hardening and sustainability of software critical
to the Nation's research; and funding for the creation of
complexity-hiding interfaces such as Science Gateways. Because
most existing software does not efficiently exploit the power
of multicore processors, and because there is insufficient
funding for development of new multicore programming tools and
applications, much of the power of multicore processors goes
unexploited in computing systems ranging from laptops to the
largest supercomputers. Because there is insufficient funding
for hardening of innovative software, and its maintenance as a
general tool for public sector research, much innovative and
useful software is not as widely (or as long) used as it could
or should be. And the lack of sufficient funding for the
development of complexity-hiding interfaces is one of the
primary factors behind the difficulty U.S. researchers have in
using current NITRD-funded cyberinfrastructure. U.S. research
competitiveness suffers as a result of all of these factors.
Increased attention to training and development of
the next generation of researchers and programmers. We note in
particular that parallel computing--one of the most difficult
forms of computing for programmers to master--has migrated from
what was once a tiny niche of the computing market into the
overwhelming majority of computing systems, from mainframes to
laptops, and soon into cell phones as well. At colleges and
universities worldwide, computer science departments are
uncovering serious deficiencies in their ability to teach
parallel computing. Because of lack of training, there is more
important work to be done than there are researchers and
programmers available in the U.S. to do it. That means that
important work is either not done at all or is done outside the
U.S.
I note that it is likely rare that any concerned individual or
group representative appears before Congress and states ``the area in
which we work receives too much money'' or even ``the area in which we
work is properly funded.'' CASC's recommendation for an increase in the
NITRD program is not intended to be self-serving, however. The areas in
which increased funding is most critically needed (expanded national
cyberinfrastructure, and even greater expansion in funding for software
and education and training) are needed so that the networking and
information technology can better serve the other U.S. communities and
research disciplines. The need is severe as well. The funding needed to
ensure U.S. leadership in networking and information technology
innovation, and the application of these innovations in ways that
maintain U.S. global leadership, is not a few percentage points but
rather on the order of hundreds of millions of dollars. We recognize
that there are many pressures on the federal budget. Maintaining U.S.
global leadership in networking and information technology is essential
to the long-term U.S. security and prosperity. Increased investment is
essential now.
Q4. The President's council of Advisors on Science and Technology
(PCAST) recommends that the NITRD program provide increased support for
research on software but does not cite specific research needs where
the current program is deficient. Do you have recommendations for how
the NITRD's investments in software research could be strengthened,
assuming no substantial increases in overall funding?
A4. Let me begin by reiterating the point made in response to an
earlier question: the objectives set forth in the PCAST recommendations
cannot be carried out effectively, in ways that preserve U.S.
international competitiveness, without substantial increases in the
NITRD budget.
Having said that, CASC suggests that the following specific areas
of software research are currently inadequate and should be
particularly strengthened:
Programming languages, compilers, run-time
environments, and performance analysis and management tools,
particularly for parallel computing using multicore processors.
Complexity-hiding interfaces, such as Science
Gateways, so that the benefits of advanced networking and
information technology systems may be more easily used by the
U.S. science and engineering communities.
Software for management and analysis of massive data
sets and real time data streams, including automated metadata
creation, provenance management, and real time analytic and
visual analysis tools.
For advanced simulation and prediction software,
increased funding for interdisciplinary research that will test
simulations and predictions against real life phenomena and
improve software accuracy and validity as a result.
Across all areas of software research, funding for
the transformation of innovative software into software that is
robust, widely usable, well supported, and maintained over
time. This recommendation echoes my testimony on the point of
consistency in funding streams over time and its importance in
preserving the human capital and expertise required to keep
software useful over time. Within the trio of research,
development, and delivery, U.S. competitiveness in the future
will be critically dependent upon more funding for development
and delivery.
Q5. Does the research community--both academe and industry--have a
voice in influencing the research priorities under the NITRD program?
Are improvements needed in the external advisory process for the NITRD
program?
A5. The Committee on Science and Technology's hearing of 31 July was an
excellent opportunity for the research community to comment on the
priorities of the NITRD program, and we very much appreciate that
opportunity.
The research community has a voice, particularly through PCAST, but
it is a voice that is less focused on networking and information
technology--and thus less strong--than during the tenure of the
President's Information Technology Advisory Committee. We thus
recommend the re-creation of a President's Information Technology
Advisory Committee to add to the advisory input that is currently heard
via the excellent work of PCAST.
Answers to Post-Hearing Questions
Responses by Don C. Winter, Vice President, Engineering and Information
Technology, Phantom Works, the Boeing Company
Questions submitted by Chairman Bart Gordon
Q1. Are there areas of IT research that are good candidates for
international cooperative efforts that would leverage U.S. investments
but would not otherwise harm U.S. competitive advantages, such as being
first to market for a new technology? What mechanisms are available, or
could be instituted, to facilitate such international cooperative
research?
A1. We believe that there are several areas that would be good
candidates. The areas of High End Computing Infrastructure and
Applications (HEC l&A). High End Computing R&D, Human Computer
Interactions and Information Management (HCI & IM), and Large Scale
Networking (LSN) are good candidates for international cooperative
efforts. They are rich in pre-competitive research challenges and offer
substantial leverage opportunities and potentially significant cost
savings. International cooperative research might be facilitated
through the creation of joint projects addressing ``grand challenges.''
Targeted expenditures for international collaboration are already in-
place at NSF--however, we are not aware of metrics that enable us to
assess the leverage or potential cost savings to development programs
of these collaborations.
Q2. Do you have suggestions for how the U.S. could institute a public-
private research partnership to advance the capabilities of cyber-
physical systems analogous to the EU's ARTEMIS initiative? Could such
an undertaking be planned and carried out under the NITRD program?
A2. We believe that this could be achieved by creating Industry/
University Consortia to perform pre-competitive research on industry-
provided testbeds. The ``industrial strength'' fidelity of the testbeds
is critical to ensuring that the research focuses on the highest
payback elements of the problem space of cyber-physical systems.
Consortia focused on more applied levels have been highly successful
and include USCAR (U.S. Council for Automotive Research) and AVSI
(Aerospace Vehicle Systems Institute).
Fundimg for the consortia could be assembled from: 1) Industry:
Internal Research & Development; 2) Academia: Government; 3) Testbed--
Government.
We propose a model based upon joint work of integrated projects as
opposed to loose/spontaneous collaborations. While the latter model can
sometimes produce important benefits, we believe the focus needs to be
the synergistic development of fundamental science directly motivated
and evaluated on realistic challenge problems from industry. In this
rapidly evolving field where time and resources are limited, this is
the most effective way to build a core technology base. Knowledge and
technology is best transitioned by people working on well defined
problems using industrial strength testbeds.
While it is possible that this activity could be carried out under
NITRD, it is not clear that there are existing mechanisms for
accomplishing this. From our perspective, a task force consisting of
representatives from industry and academia should be created to examine
potential models, and recommend the appropriate structure within the
next 90 days. A further recommendation is that task force leadership
should be provided by industry to emphasize the need to consider novel
organizational and execution models.
Q3. Are the R&D objectives of the NITRD program, as it is currently
constituted, prioritized appropriately and is the allocation of funding
consistent with achieving the objectives? Are there particular research
areas that the NITRD program is not pursuing with sufficient resources?
A3. From our perspective the R&D objectives, as indicated by funding
levels, are not optimally prioritized. Nearly 50 percent of the FY 2008
and 2009 NITRD budgets ($1.5B out of $3.3B in FY08) are allocated to
High End Computing (Architecture, Infrastructure, and R&D). HEC is not
at present an area where we feel U.S. competitiveness is at stake.
Expenditures for Human Computer Interaction and Information Management
($0.8B) also appear out of proportion relative to the need and
potential gains in research and competitiveness to be attained. The
PCAST correctly pointed out the need to substantially increase the
level of spending on CPS--which is not even explicitly mentioned among
the programs in NITRD budget documents.
We believe that CPS should be included as a separate program under
NITRD since this would provide transparency of budget allocations to
this critical technology area. Fundamental research progress in CPS
will have long-term benefits to assuring competitive position in
Aerospace, Automotive, Energy Management, Health and other areas.
Furthermore, the low level of funding for High Confidence Software and
Systems (HCSS) ($0.12B), and Software Design & Productivity (SDP)
($0.073B) do not adequately address the needs and challenges for CPS,
and are insufficient to stimulate breakthroughs required, especially in
large scale CPS systems common to DOD and commercial Aerospace and
Energy applications.
We are also concerned about the isolation of Cyber Security and
Information Assurance (CSIA) from the systems domains (HCI & IM, LSN,
HCSS, SEW, SDP). CPS must include an essential CSIA program element
because of the unique vulnerabilities and consequences associated with
the target industries. What we need is CPS focused R&D in CSIA, tightly
integrated with all other research challenges.
Q4. The President's Council of Advisors on Science and Technology
(PCAST) recommends that the NITRD program provide increased support for
research on software but does not specify research needs where the
current program is deficient. Do you have recommendations for how the
NITRD's investment in software research could he strengthened, assuming
no substantial increases in overall funding?
A4. A number of workshops have been conducted since 2005, led by Dr.
Andre Van Tilborg (Director of the Information Systems Directorate in
the Office of the Deputy Under Secretary of Defense for Science and
Technology) that sought Government, Industry, and Academic perspectives
on Software Intensive Systems. The workshops highlighted the clear need
and established a framework for a software research agenda. The
aerospace industry in particular (Boeing, Lockheed, Raytheon, Northrop,
BAE, Honeywell, among others) presented a common perspective on the
need and benefits of testbed driven software research. Elements of the
research agenda are published in a study on Ultra Large Scale Systems.
While the total level of NITRD investment may be adequate, funds should
be shifted from HEC and HCI & IM to CPS focused investments in HCSS,
Cyber Security, and Software Design due to its significantly larger
impact on U.S. competitiveness. In addition, increased investment in
these areas holds significant potential for reducing costs for DOD CPS
software developments.
Q5. Does the research community--both academe and industry--have a
voice in influencing the research priorities under the NITRD program?
Are improvements needed in the external advisory process for the NITRD
program?
A5. Industry has had a very limited voice in influencing research
priorities of NITRD program. Organizations like PCAST have influence at
a strategic level but they have little influence in implementation. We
believe that proactive industrial participation in shaping NITRD
priorities and participation in the research agenda is key to achieving
breakthroughs required.
Q6. The top recommendation of the PCAST assessment of the NITRD
program for new research investments is in the area of cyber-physical
systems. You have been engaged with your colleagues from industry,
academia, and government in discussions of research requirements in
this field. What is the status of these discussions? Are there plans to
develop a set of research goals and a roadmap for achieving these
goals?
A6. The discussions on research requirements continue. Under the
auspices of the aforementioned workshops, we have formed industry teams
in aerospace, energy management, and automotive and are working to
finalize technology roadmaps. The discussions were initiated in May
2007 at a CPS roundtable conducted with representatives from industry
(BAE, Boeing, Lockheed, Honeywell, United Technology, IBM, etc.).
Academia (Carnegie Mellon, UCB, Vanderbilt, etc.), and Government (NSF,
DOD, etc.). We have developed initial industrial roadmaps that need to
be finalized and then shared with the research community at large.
Q7. What do you see as the relative roles of industry-sponsored
research and federally-sponsored research in moving this technology
area forward and giving the U.S. a strong position?
A7. The major issue here is that the CPS research agenda is cross-
cutting and spans multiple industries. Much of the research required is
of a pre-competitive nature--where industry-sponsored research dollars
are inherently limited. The current approach of Federal Government-
sponsored research in this area has not adequately addressed
``industrial strength'' real-world challenge problems, and not created
significant transition pathways outside of the academic world. Greater
industrial participation in executing the research agenda is critical
to success and will spur the focused industrial-academic collaboration
needed for significant progress.