Ask the Scholar
Document scope · 1 page
Scholar
Ask about this object, its catalog metadata, its source description, or the page inventory.
For page-specific OCR and visual context, open one of the page chats.
Source Description
Records pertain to the Office of Science and Technology Policy.
Scholar Source Context
Document identity
localId
285790773
label
Science & Technology [4]
core
doc
dtoType
document
citationUrl
pageCount
1
Source metadata
id
285790773
contentType
document
title
Science & Technology [4]
description
Records pertain to the Office of Science and Technology Policy.
citationUrl
identifierLocal
04296-004
collections
Records of the Economic Policy Council (George H. W. Bush Administration)
Olin Lewis Wethington Subject Files
imageCount
1
hasImages
yes
source
import
hasTranscription
no
Source extras
naId
285790773
levelOfDescription
fileUnit
recordType
description
ocrSource
nara-archive
Single page context
seq
1
pageIndex
0
type
document
mediaId
1e862e7ef5a50b23
ocrText
Originally Processed With FOIA(s):
FOIA Number:
2005-0336-F
2005-0336-F
FOIA
MARKER
This is not a textual record. This is used as an
administrative marker by the George Bush Presidential
Library Staff.
Record Group/Collection:
George H.W. Bush Presidential Records
Collection/Office of Origin: Economic Policy Council
Series:
Wethington, Olin, Files
Subseries:
Subject Files
OA/ID Number:
04296
Folder ID Number:
04296-004
Folder Title:
Science & Technology [4]
Stack:
Row:
Section:
Shelf:
Position:
G
13
28
4
2
A
1
CONTENTS
2
STATEMENT OF:
PAGE
3
HON. ALBERT GORE, JR., U.S. SENATOR FROM TENNESSEE
2
4
HON. LARRY PRESSLER, U.S. SENATOR FROM SOUTH DAKOTA
12
5
HON. ROBERT W. KASTEN, JR., U.S. SENATOR FROM WISCONSIN
20
6
D. ALLEN BROMLEY, DIRECTOR, OFFICE OF SCIENCE
25
7
AND TECHNOLOGY POLICY
8
HON. LARRY PRESSLER, U.S. SENATOR FROM SOUTH DAKOTA
47
9
HON. CHARLES S. ROBB, U.S. SENATOR FROM VIRGINIA
54
10
DONALD LANGENBERG, CHANCELLOR OF THE UNIVERSITY OF
67
11
MARYLAND SYSTEM AT ADELPHI, MARYLAND
12
MALVIN H. KALOS, DIRECTOR, CORNELL THEORY CENTER
74
13
MR. TRACEY GRAY, VICE PRESIDENT OF MARKETING,
83
14
GOVERNMENT SYSTEMS DIVISION, US SPRINT
15
COMMUNICATIONS COMPANY LIMITED PARTNERSHIP
16
DR.. DAVID C. NAGEL, VICE PRESIDENT, ADVANCED
89
17
TECHNOLOGY, APPLE COMPUTER, INC.
18
DR. JOHN S. WOLD, EXECUTIVE DIRECTOR, LILLY RESEARCH
99
19
LABORATORIES, ELI LILLY AND COMPANY; ACCOMPANIED BY
20
DR. RIAZ ABDULLA, MANAGER, HIGH-PERFORMANCE COMPUTING
21
PROGRAM, ELI LILLY AND COMPANY
22
23
)
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
1
1
HEARING ON
2
S. 272
3
- HIGH PERFORMANCE COMPUTING AND COMMUNICATIONS
4
5
Tuesday, March 5, 1991
6
7
U.S. Senate
8
Subcommittee on Science,
9
Technology, and Space
10
Committee on Commerce, Science,
11
and Transportation
12
Washington, D.C.
13
The subcommittee met, pursuant to notice, at 2:04 p.m.,
14
in Room SR-253, Russell Senate Office Building, Hon. Albert
15
Gore, Jr., presiding.
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
2
1
OPENING STATEMENT OF HON. ALBERT GORE, JR., U.S. SENATOR
2
FROM TENNESSEE
3
Senator Gore: This subcommittee will come to order.
4
I would like to welcome Dr. Bromley and Dr. Wong and
5
distinguished guests, other witnesses, and ladies and
6
gentlemen.
7
I will have an opening statement, and then I will call on
8
Senator Pressler, and then we will move right into the
9
testimony of Dr. Bromley.
10
Today the Science and Technology Subcommittee is
11
considering S. 272, the High Performance Computing Act. This
12
bill is designed to ensure that the United States stays at the
13
leading edge in computer technology. It would roughly double
14
the Federal investment in research and development in new
15
supercomputers, more advanced software, and high-speed
16
computer networks.
17
Perhaps most importantly it would create a national
18
research and education network, the NREN -- or the National
19
Information Superhighway -- as I like to call it, which would
20
connect more than 1 million people at more than 1 thousand
21
colleges, universities, laboratories, and hospitals throughout
22
the country, giving them access to computing power and
23
information -- resources unavailable anywhere today -- and
24
making possible the rapid proliferation of a truly Nation-
25
wide, ubiquitous network which can do more to enhance our
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
3
1
Nation's productivity than any other single development.
2
These technologies and this network represent our
3
economic future. They are the smokestack industries of
4
today's information age. We talk a lot now about jobs and
5
economic development, pulling our country out of recession and
6
into renewal.
7
Well, our ability to meet the economic challenges of the
8
information age and beyond -- tough challenges from real
9
competitors around the world -- will rest in large measure on
10
our ability to maintain and strengthen an already threatened
11
lead in these technologies and industries.
12
We are witnessing the emergence of a much-heralded,
13
prematurely recognized -- on several occasions -- global
14
civilization. Now it is really here. And those nations best
15
able to deal with information will be the nations most
16
successful in this global civilization.
17
It is based on shared knowledge in the form of digital
18
code. And our ability to compete will depend on our ability
19
to handle knowledge in that form. It is now the lingua franca
20
of global civilization.
21
We used to think of our ability to compete in terms of
22
infrastructure -- did we have enough deep-water ports? Did we
23
have enough railroad lines or highways? Now we need to think
24
about information infrastructure.
25
I have been advocating legislation such as this for more
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
4
1
than 1 dozen years. Because I strongly believe it is critical
2
for our country to develop the best scientists, the best
3
science, the fastest, most powerful computers, and the best
4
base of knowledge -- and then to ensure access to these
5
technologies to as many people as possible, so as many people
6
as possible will benefit from them.
7
This legislation will help us do just that.
8
Every year there are new advocates. This year, finally,
9
President Bush is among them, including in his budget for
10
Fiscal 1992 $149 million in new funding to support these
11
technologies.
12
We cannot afford to wait or to put off this
13
challenge -- not if we care about jobs, economic development
14
or our ability to hold our own in world markets.
15
During the last 30 years, computer technology has
16
improved exponentially, faster than technology in any other
17
field. Computers just keep getting faster, more powerful and
18
more inexpensive.
19
According to one expert, if automobile technology had
20
improved as much as computer technology in recent years, a
21
1991 Cadillac would now cruise at 20,000 m.p.h, get 5,000
22
miles to a gallon, and cost only 3 cents.
23
[Laughter. ]
24
When my friend Jim Schlesinger heard someone deliver that
25
cliche recently, he said yes, and your Cadillac would be a
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
5
1
few millimeters long, too.
2
[Laughter.]
3
But as a result of these amazing advances, computers have
4
gone from being expensive, esoteric, research tools, isolated
5
in the laboratory, to being an integral part of our every-day
6
life. We rely on computers at the supermarket, at the bank,
7
in the office, and in our schools. They make our life easier
8
and better in hundreds of ways.
9
And yet, the computer revolution is far from over. In
10
fact, according to some measures, the price/performance ratio
11
of computers is improving even faster now than it has in the
12
past. Anyone who has seen a supercomputer in action today,
13
has a sense of what computers can do for all of us in the
14
future.
15
Today, scientists and engineers are using supercomputers
16
to design better airplanes, understand global warming, find
17
oil fields, and discover safer, more effective medications.
18
In many cases, they can use these machines to mimic
19
experiments that would be too expensive or downright
20
impossible in real life. With the supercomputer model,
21
engineers at Ford can simulate auto crashes and test new
22
safety features for a fraction of the cost and in a fraction
23
of the time it would take to really crash an automobile. And
24
they can observe many more variables in much more detail than
25
they could with a real test.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
6
1
The bill we are considering today is very similar to the
2
first title of S. 1067, the High Performance Computing Act of
3
1990, which passed the Senate unanimously last October.
4
Unfortunately, of course, the House was unable to act on the
5
bill before we adjourned because of differences in measures
6
passed by the two bodies.
7
It is my hope that we will be able to move this bill
8
quickly this year. There is widespread support in both the
9
House and the Senate. In the House, Congressman George Brown,
10
the new Chairman of the House Committee on Science, Space, and
11
Technology, has introduced a very similar bill -- H.R.
12
656 -- co-sponsored by Congressman Tim Valentine, Sherwood
13
Boehlert, Norm Mineta and others.
14
On Thursday, the Science Committee's Subcommittee on
15
Science and its subcommittee on Technology and Competitiveness
16
on the House side will be holding a joint hearing on the bill.
17
And I look forward to working with my House colleagues to move
18
this bill as quickly as possible.
19
This legislation provides for a multi-agency, high-
20
performance computing research and development program to be
21
coordinated by the White House Office of Science and
22
Technology policy, whose director, Dr. Alan Bromley, is our
23
first witness today.
24
The primary agencies involved are the National Science
25
Foundation, the Defense Advance Research Projects Agency, the
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
7
1
National Aeronautics and Space Administration, and the
2
Department of Energy. Each of these agencies has experience
3
in developing and using high-performance computing technology.
4
S. 272 will provide for a well-planned, well-coordinated
5
research program to effectively utilize the talents and
6
resources available throughout the Federal research agencies.
7
In addition to the agencies I just mentioned, it will involve
8
also the Department of Commerce -- in particular, the National
9
Institute of Standards and Technology, and NOAA, the
10
Department of Health and Human Services, the Department of
11
Education, the U.S. Geological Survey, the Department of
12
Agriculture, the Environmental Protection Agency, and the
13
Library of Congress as well.
14
The technology developed under this program will find
15
application throughout the Federal Government and throughout
16
the country. S. 272 will double funding for high-performance
17
computing at NSF and NASA during the next 5 years. Additional
18
funding, more than $1 billion during the next 5 years, will
19
also be needed to expand research and development programs at
20
DARPA and DOE.
21
Last year, I worked closely with Senators Johnston and
22
Domenici on the Energy Committee, to pass legislation
23
authorizing a DOE high-performance computing program. And I
24
hope to work with them and the other members of the Energy
25
Committee to see that program authorized and funded in FY
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
8
1
1991. Already, Senator Johnston and others have introduced S.
2
343, authorizing -- which would authorize DOE's part of this
3
multi-agency program.
4
To fund DOD's part of the program, last year I worked
5
with Senators Nunn and Bingaman and others on the Armed
6
Services Committee to authorize and appropriate an additional
7
$20 million for DARPA's high-performance computing program,
8
money that has been put to good use developing more powerful
9
supercomputers and faster computer networks.
10
Advanced computer technology was a key ingredient in
11
Operation Desert Storm. But we cannot simply rely on existing
12
technology. We must make the investment needed to stay at the
13
leading edge. It is important to remember that the Patriot
14
Missile and the Tomahawk Cruise missile rely on computers
15
based on technologies developed through Federal computer
16
research programs in the 1970's.
17
The High Performance Computing Act will help ensure the
18
technological lead in weaponry that has helped us win the war
19
with Iraq, and that will improve our national security in the
20
future.
21
This same technology is improving our economic security
22
by helping American scientists and engineers develop new
23
products and processes to keep the U.S. competitive in world
24
markets. Supercomputers can dramatically reduce the time it
25
takes to design and test a new product -- whether an airplane,
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
9
1
a new drug or an aluminum can. That means more energy-
2
efficient, cheaper products. It means higher profits and more
3
jobs for Americans.
4
But perhaps the most important contribution this bill
5
will make to our economic security is the National Research
6
and Education Network, the cornerstone of the program, funded
7
by the bill.
8
In 1996, this fiber optic computer network will connect
9
more than 1 million people at more than 1 thousand colleges
10
and universities in all 50 states, allowing them to not only
11
send electronic mail and shared data, but access
12
supercomputers and use research facilities such as radio
13
telescopes; log-on to databases containing trillions of bytes
14
of information on all sorts of topics.
15
This network will speed research and accelerate
16
technology transfer so that the discoveries made in our
17
university laboratories can be quickly and effectively turned
18
into profits for American companies.
19
Today the National Science Foundation runs NSF NSFNET,
20
allowing researches and educators to exchange up to 1.5 bytes
21
of data per second. The NREN will be at least 1,000 times
22
faster, allowing researchers to transfer all the information
23
in the entire Encyclopedia Britannica from coast to coast in
24
seconds.
25
With today's networks, it is easy to send documents and
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
10
1
data, but images and pictures require much faster speeds.
2
That will require the NREN, which can carry billions of bytes
3
per second.
4
That is important. Because one of the only ways we can
5
successfully deal with the mountains of excess data we now
6
have is by organizing it into coherent, mosaic patterns and
7
images which can be comprehended in gulps or chunks, instead
8
of byte-by-byte, one byte at a time. It is impossible to deal
9
with this much information in any other way.
10
With access to computer graphics, researchers throughout
11
the country will be able to work together far more effectively
12
than today. It will be much easier for teams of researchers
13
at colleges throughout America to work together. They will be
14
able to see the results of their experiments as the data comes
15
in. They will be able to share the results of their computer
16
models in real time, and brainstorm by tele-conference.
17
William Wulf, formerly Assistant Director for Computer
18
and Information Science at NSF likes to talk about the
19
"National Collaboratory" -- a laboratory without walls, which
20
this network will make possible.
21
Researchers throughout the country at colleges and labs,
22
large and small, will be able to stay on top of the latest
23
advances in their fields. The NREN and the other technology
24
funded by S. 272 will also provide enormous benefits to
25
American education. And I will include in the record a
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
11
1
lengthy description of how education will be improved by this
2
network.
3
In conclusion, let me say that with this bill, I believe
4
we can help shape the future and shape it for the better.
5
This is an investment in our national security and our
6
economic security which we cannot afford not to make. For
7
that reason, I was very glad to see the administration propose
8
a high-performance computing and communications initiative, a
9
program very similar to the one outlined in S. 272.
10
In intent to work closely with Dr. Bromley and others
11
within the administration, as well as my colleagues in
12
Congress, to secure the funding needed to implement this
13
critically important program.
14
Before turning to Dr. Bromley, I would recognize now,
15
Senator Pressler. I apologize for the length of my opening
16
statement. But after 12 years, I have a lot to say about this
17
topic. And I will now recognize my friend, the Senator from
18
South Dakota.
19
[The prepared statement of Senator Gore follows:]
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
12
1
STATEMENT OF HON. LARRY PRESSLER, U.S. SENATOR FROM
2
SOUTH DAKOTA
3
Senator Pressler: Thank you, Mr. Chairman.
4
And I think it is very appropriate that we have Dr.
5
Bromley here at this opening hearing. I want to thank you,
6
Mr. Chairman, for holding this hearing on S. 272, and for your
7
early and continued leadership in support of a national
8
supercomputer network.
9
Let me say that I am glad that the administration has
10
come with the proposal this year. I challenged the
11
administration to come with a proposal last year. I am a
12
cosponsor of this legislation. But I look forward also to
13
working out the differences and listening very closely to the
14
administration, here today.
15
Let me -- rather than repeat everything that Al has said
16
in terms of what supercomputers can do -- to cite a special
17
interest that I have. And that is to be sure that small
18
businesses, smaller cities and towns, smaller universities,
19
Indian reservations, and others are taken care of this
20
communications transformation we are going through.
21
I am very interested, for example, in finding a way that
22
fiber optics cable can be laid to every household and business
23
in the United States, and not leave some out. Generally
24
speaking, in the Communications Subcommittee and this
25
Subcommittee, when people talk about supercemputers or fiber
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
13
1
optics cable, or new electronic services, they are talking
2
about serviceAthe to wealthy suburbs. They are talking about
3
serving densely populated areas where there is a market that
4
yields greatest profits.
5
I always cite as an example that in cable T.V., for
6
example -- my wife and I have just obtained cable T.V. at one
7
of our homes for the first time of last year, we live in
8
Washington, D.C. where we have a home, and Humbolt, South
9
Dakota. In neither place could we get cable T.V. until this
10
past year. We finally have it in Washington, D.C. And we can
11
get it by microwave now, in Humbolt, South Dakota.
12
But the point is that the rural and smaller town areas
SMALL SCHO
13
have something in common with inner cities, in that all the
they are not TALKING about ^
A merica
14
great talk about the communications revolution, people they
are
15
talking about the largest universities, the wealthiest
16
suburbs, and so forth. And We cannot become a Nation of two
17
or three communication systems.
18
x I am working on legislation in the Communications
19
Subcommittee related to this that would provide that the
20
telephone companies, or whoever, or the cable companies should
21
be laying fiber optics cable rather than copper cable. That
22
we should get fiber optics to everybody. And from there we
23
can rent space on the fiber optics cable for the different
determine how
24
users or figure it out somehow 30 these services can be
POLICY.
25
available on a national basis sort a of universal service so-
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
14
1
to speak.
2
And the same is true for supercomputers. Of course,
3
every small business does not need a supercomputer. But many
4
of our universities do -- many small businesses will benefit
5
from being hooked-up by fiber optics cable. In my own State
6
there are certain institutions. For example, at South Dakota
7
State University, the bio-stress lab where they are studying
8
new kinds of plants, resistance to drought, very much wants to
9
be connected to a supercomputer.
Archives
10
The EROS Data Center, which has pictures
the LANDSAT
researchers
11
pictures, very much will benefit, and the country will
the EROS DATABASE
12
benefit from being hooked to a supercomputer.
13
South Dakota School of Mines, the project such as the
14
deep-drilling project, certain EPSCoR-type projects, where
15
professors and individuals are doing research, and they are
16
not associated with a huge university will benefit a great
17
deal.
18
Indeed, you can build the argument that small hospitals
19
and small universities need this
need access to a
20
supercomputer more than a big one. Because the big ones
21
probably have their own. So I think this is a very important
22
initiative.
23
Let me also say small business -- now not every small
24
business is going to want to be hooked to a supercomputer.
25
But if they have the option, as we move into the 1990's, there
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
15
1
are jobs to be created in areas -- everybody does not have to
2
be in the Boston beltway or a place like that to do technical
3
work. And it will enrich our country. Now I am not saying
4
there is anything wrong with being in the Boston beltway area.
5
But we also need a few in the Sioux Falls beltway area, who
6
have access to a supercomputer.
7
[Laughter.]
8
I am also concerned, Mr. Chairman, with -- in addition to
9
this, and in this subcommittee -- I hope we look at this year
10
what is happening in our schools in math and science training.
11
I think we are falling behind in science and mathematics.
12
Someone who is 15 or 18 years old and has not gotten the
13
basics is out of the system forever. Where our educational
14
system works, I am told that over half of our graduate
15
students in engineering are from abroad. Part of that is due
16
to our pay structure. The thing to be is an investment banker
17
or a lawyer and move papers that result in no productivity but
18
the pay is pretty good.
We have to think very hard about
19
the impact, the long-term impact of losing the math and
20
science edge.
21
There are a whole series of other things. I was going to
22
go into some of the benefits of supercomputers. The Chairman
23
has done that very aptly in his statement. I have mention
24
South Dakota State University's bio-tress laboratories is
25
developing strains of crops and livestock. With the
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
16
1
supercomputer network, the results of the researched produced
2
in the bio-stress labs could be placed on-line for ready
3
access_ for users at other universities that are doing similar
4
research.
5
And let me say that I think many of our smaller
6
universities very much want to be connected to some of the
7
work at larger universities -- certainly some of our smaller
8
hospitals very much are in need of this type of service.
9
I have mentioned the EROS data center archives -- the
10
pictures that are sent back from our LANDSAT
11
satellite -- since 1971, near Sioux Falls, it has stored over
12
1 million satellite images on magnetic tape. Government and
13
university researchers and commercial enterprises use the
14
LANDSAT images for oil and mineral exploration, land use
15
planning, map making, climate change monitoring, crop
16
assessments, and many other applications.
17
With the proposed network, any researcher could use the
18
LANDSAT pictures that are stored in South Dakota without
19
having to leave his or her desk or laboratory. This
20
capability will become even more important when the EROS data
21
center is used later in this decade to achieve the massive
22
amounts of data for the Earth Observing System, EOA, NASA's
23
contribution to our multi-agency global change research
24
program. And I was proud to be on hand last summer when we
25
raised the NASA flag at the EROS data center in Sioux Falls.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
17
1
The EOS satellites will be transmitting the data
2
equivalent of the entire Library of Congress every 5 days. I
3
think that is really an amazing thing. To use this data
4
effectively, scientists will need ready access to the EROS
5
Data Center database.
6
Mr. Chairman, I also want to take this time to commend
7
President Bush for his support for our multi-agency
8
supercomputing initiative in its Fiscal Year 1992 fiscal year
9
budget request. President Bush has allocated $638 million to
10
a supercomputer network and research effort. The Bush plan,
11
which has the same purposes as S. 272, is patterned after a
12
program set forth in a 1989 report by the White House Office
13
of Science and Technology Policy. As I have mentioned, I am
14
pleased that Dr. Alan Bromley, the Director of that office is
15
testifying today. I look forward to hearing his details about
16
the administration's initiative.
17
I am hopeful that during this Congress the administration
18
and congressional supporters of supercomputing will join
19
forces to implement a national, multi-agency supercomputing
20
initiative -- boy that is a mouthful of Washington
21
words -- whether done by statute or otherwise.
22
Mr. Chairman, I am sure we will all agree that the main
23
goal here is to begin in this Congress a national commitment
24
to supercomputing and a national supercomputing network
25
linking more than 1 million of our country's computers by
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
18
1
1996.
2
Mr. Chairman, I look forward to hearing from our
3
distinguished panel of witnesses on both S. 272 and
4
supercomputing generally.
5
[The prepared statement of Senator Pressler follows:]
6
LEOMMITTEE INSERT]
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
19
1
Senator Gore: Thank you, Senator Pressler.
2
Senator Kasten.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
20
1
STATEMENT OF HON. ROBERT W. KASTEN, JR., U.S. SENATOR
2
FROM WISCONSIN
3
Senator Kasten: Mr. Chairman, I am not going to try to
4
compete with you and Senator Pressler in terms of length of
5
opening statements.
6
So what I would like to do is ask unanimous consent that
7
the entire statement that I have appear in the record and that
8
I be allowed to excerpt a couple of quick points from it.
9
I first of all want to thank you for the hearing. And I
10
am pleased to be an original co-sponsor of this piece of
11
legislation. I agree with both of you that it is necessary
12
for the continued ability to innovate and compete in the
13
marketplace.
14
I am also pleased that though there are differences in
15
emphasis, the administration recognizes the needs for the
16
kinds of programs that we are supporting in S. 272. I believe
17
that the legislative initiatives, particularly S. 1067, in the
18
last Congress, Mr. Chairman, some of these initiatives and
19
other pressures from Congress is, in fact, what has helped the
20
administration -- helped assure the Administratif's attention
21
to this important area. And for this I commend you.
22
And I also commend you for you statement -- in your
23
opening statement saying that we want to work with the
24
administration. Because I believe that we are close on the
25
dollars and not SO far apart on the programs. And SO we have
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
21
1
the opportunity here now to come together. But I think it was
2
Congressional initiative, and particularly your leadership,
3
Mr. Chairman.
4
The future developments in the area of supercomputing,
5
hardware, software, and communications, are critical to the
6
country's security -- economically, environmentally, and
7
militarily. It is important that an American company -- it
8
also happens to be a Wisconsin company, Cray Research, is the
9
current world leader in the field. And I believe that our
10
bill will help to keep American supercomputer companies
11
healthy. And it will also assure that American companies,
12
researchers, and others who use supercomputers have the
13
software, the training and the access so that we are all using
14
our capabilities and using them to the fullest.
15
Both S. 272 and the administration's proposal look to
16
develop the networks that will be necessary if there is going
17
to be broader access and broader utilization of the Nation's
18
supercomputers. Through the computing capability of our
19
supercomputers -- although it is incredible -- our
20
communications systems are not up to the task of transmitting
21
the-data between computers or to -- as Senator Pressler
22
discussed. -- some remote locations.
23
So I think we are on the right track here. I will work
24
to assure that America realizes the full promise that
25
supercomputers can offer in SO many different areas of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
22
1
scientific, educational, and commercial endeavor. And I think
2
that we, Mr. Chairman, have not only a hearing on a very
3
important subject today, but we have the beginnings now of a
4
mark-up of a bill that is going to be passed-out by the
5
Committee, going to be passed by the Senate, passed by the
6
House of Representatives, and, in fact, signed into law.
7
And that is, I know, your goal and your purpose. And I
8
just want to say that I am an anxious lieutenant on this team
9
or in this area that we are going to do something here. And
10
it is more than just simply a hearing on a very important
11
subject. And we all will work together. And I look forward,
12
Mr. Chairman, to working with you.
13
Unfortunately, I have another hearing that also started
14
at 2:00. And I would ask that in addition to my entire
15
statement appearing in the record, that I be permitted to
16
submit some questions for the record of Dr. Bromley.
17
Senator Gore: Absolutely, without objection.
18
Senator Kasten: Thank you, Mr. Chairman.
19
[The prepared statement of Senator Kasten follows:]
20
[COMMITTEE INSERT]
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
23
1
Senator Gore: And may I say, before you must depart, on
2
a personal note, I deeply appreciate the key role that you
3
have played in pushing this bi-partisan effort forward. And I
4
want to say the same to Senator Pressler.
5
Indeed, this has been bi-partisan from the very
6
beginning. And you mentioned the cooperative relationship
7
between the committee and the administration. That is a
8
spirit we want to continue.
9
In fact, the OSTP plan which Senator Pressler mentioned
10
coming out in 1989, really came about because of legislation
11
this committee passed which both of you supported, the
12
Supercomputer Network Study Act of 1986 which required that
13
study, but more than requiring it, invited the administration
14
to join in a dialogue with the committee which resulted in the
15
OSTP plan 2 years ago.
16
And it is not accidental, by any means, that the
17
administration plan and the legislation which we here are
18
supporting, are so similar. Because they have both resulted
19
from a meeting of the minds on what is the best interests of
20
our country.
21
Dr. Bromley, again, let me apologize for the length of
22
time it has taken us to get to your statement. We are very
23
interested in it.
24
I want to acknowledge the presence of Dr. Eugene Wong,
25
Associated Director for Physical Sciences at OSTP. And among
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
24
1
the many things for which you are due congratulations, is the
2
excellent choices you have made of people to help you.
3
I-remember when you first took the post, you said you
4
were going to do it. I agreed with your assessment that this
5
was one of the keys to making this office a more integral part
6
of the Government. And congratulations on getting the
7
President on board on this initiative.
8
I know you would probably want me to choose different
9
words in saying that, but that is what I feel about it. I
10
think this is you speaking for the administration now. And we
11
welcome you. So please proceed.
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
25
1
STATEMENT OF D. ALLEN BROMLEY, DIRECTOR, OFFICE OF
2
SCIENCE AND TECHNOLOGY POLICY
3
Dr.-Bromley: Thank you, Mr. Chairman.
4
I welcome this opportunity to appear before you and the
5
members of the committee to discuss this initiative. And I
6
would, in fact, begin by acknowledging the leadership that you
7
and this group have shown in this area that I agree is one of
8
the most important that we could address in terms of its long-
9
term and short-term impact on this Nation.
10
The hearing, of course, today addresses this whole
11
question of high-performance computing and communication. And
12
I would like to, with your permission, sir, have the formal
13
testimony that I have provided to the committee included in
14
the record and simply abstract it.
15
Senator Gore: Without objection, that would be fine.
16
Dr. Bromley: Thank you, sir.
17
Now, the President's initiative included in the 1992
18
budget is described in detail in the document that accompanies
19
that, precisely, the report that is called the Grand
20
Challenges: High Performance Computing in Communication.
21
This is a report that was prepared by a working group on high-
22
performance computing and communications that operates under
23
the Committee on Physical Sciences and Engineering, one of the
24
seven umbrella committees that operates under the Federal
25
Coordinating Council for Science, Engineering and Technology.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
26
1
And L I see, sir, that you already have a copy of the report.
2
The goals of the whole initiative high performance
3
computing and communication I think, sir, are ones that we
4
have shared for a long time. And my staff and I look forward
5
to working with you to realize these goals for the American
6
people.
7
In the document that we have provided to you, the Grand e
8
Challenges document, we have attempted to illustrate and
9
symbolize the importance of this activity by setting forth a
10
series of very important scientific and societal problems
11
whose solutions simply elude us at the moment, but would
12
become within reach with the kind of program that both of us
13
are talking about.
the questions
14
These include matters of global climate change, mapping
15
the human genome, understanding the nature and helping in the
16
fabrication of tailored materials --
an entirely new
and
17
frontier problems that are directly applicable to our
18
19
would agree with L us sir -- both military strength and economic
national security, and that includes /ing -- and I believe you
20
strength. It also, of course, includes the design of ever
21
more sophisticated computers. And Essentially, I think that the list
22
essentially I. is limited only by our imagination at the moment e
23
A great many of these topics of course, I think are
24
already partially underway. But there is a tremendous
promising
25
frontier out there, and qualitative as well as quantitative
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
27
1
changes in our capabilities and in the problems that we can
2
address effectively.
3
The initiative that we present to you represents a full
4
integration of component programs I in a whole number series of the l
5
Federal agencies These that have agencies been involved for some time I sir and
6
in high Fe performance computing and in the design, development,
7
utilization or { their own agency computer networks. And as was
8
noted earlier, the initiative that the President has brought
9
forward proposes a 30 percent increase in the funding in
to
10
Fiscal Year 1992 in n support of these activities.
11
And ₦ think that I should la at the outset S SIT, state that
12
it is our goal in OSTP that between the next 4 and 5 years we
will be doubled But it is
13
double the support for this activity but much more important
14
that we increase the speed, the memory capacity, of and the data
15
transmission capacity of our systems by factors between 100
16
and 1000. And this is certainly within technical reach.
now
17
Now, I would like, if I might, sir, to take a moment just
18
to trace the history of this initiative, in which you have
19
been, yourself, very heavily involved. I think that it traces
20
its formative years back to the early 1980's and before, and
21
stems also & & in part from the recognition in a great many of our
22
agencies I that to satisfy their mission needs I they required
23
computing capacity far beyond anything that was then chenf
24
available.
25
And as the science and the technology in these agencies
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
28
and
1
improved, it became rapidly obvious that the quantity of data 6"
2
the number of databases that were being developed made
3
absolutely mandatory the need and for more sophisticated data-
4
archiving, data-processing,, data retrieval. And one of the
5
points I would make in passing sir, I is that there are now
6
between 6,000 and 7,000 data bases in use around the world.
7
And & almost without exception, these are not inter I comparable
8
or readable, one by any other. And this is the beginning of a
9
Tower of Babel that we can ill afford. And that matter is one
10
that I think that we should work together to correct.
11
In 1982 I a FCCSET committee examined the status of
12
supercomputing in the United States in response, as you
13
suggested, to a request from the Congress and reviewed the
14
role of the Federal Government in this area of technology.
15
In 1985 the committee recommended Government action
16
necessary to sustain the technological superiority that we had
17
at that time, and to further the development and use of
18
supercomputers in this country. Subsequent planning resulted
19
in a who 1e series of workshops that were held in 1987 la and a
a
20
set of reports that set forth the outlines of, research and
21
development strategy. The synthesis of all of this activity
22
appeared in the report entitled the "Federal High Performance
23
Computing Program" that was issued by my office in September
24
of of & 1989, as you noted. And the initiative that we bring
25
forward now in the 1992 budget is, in my view, a realization
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
and
29
1
of the goals that were spelled-out by you, by us, by a number
of people back in those days as being thengoals, but now & they
they were seen as
2
are
3
^ targets that we fully believe The we can achieve.
4
What we have here is / an program that involves 8 partners
S
5
And this is the Defense Advanced Research Project Agency, the
6
National Science Foundation, the Department of Energy,
the
Administration, the)
7
A
National Aeronautics and Space He National Library of Medicine,
8
the Environmental Protection Agency, the National Institute of
9
Standards and Technology, and NOAA, the National Oceanic and
10
Atmospheric Administration, within Commerce.
11
And And the planning and implementation of the HPCC program,
12
I would submit to you, sin have been the result of a rather e
13
remarkable degree of cooperation and I must say Q
14
enthusiastic cooperation wherein I these and agencies have been
15
quite prepared to readjust, to I realign, redesign their
in
Y
16
programs and high-performance computing so that together these
17
made parts of a coordinated Federal whole.
18
And & I would like to pay tribute to the level of
19
cooperation that we have enjoyed in putting this program
20
together. There is a level of mutual trust, cooperation, and
21
synergism It that I believe & is remarkable / both inside and
22
outside of Government. And I would also have to say that the
23
success of this activity has depended in no small measure on
24
the input that we have received from the private sector. It
25
has been crucially important for us in developing the program
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
30
and
1
that we have brought before you to have it calibrated I to have
2
it A tested against the real-world environment in which our
3
industries operate, because they, in the long-run, are the
4
ones that we look to to implement this program.
5
The program itself has four major components -- as you
concerns
6
know very well, sir. The first has to do with high-
a
7
performance computing systems. And we use that phrase
8
advisedly because we wanted to include not just
9
supercomputers of but also high-performance computers on all
10
levels because the synergism there is important and should
11
not be forgotten.
12
Secondly, we talk about advanced software and algorithms,
13
because without that, all the hardware in the word is
14
essentially useless.
15
Thirdly, we have the National Research and Education
16
Network, the information superhighway to which you referred.
17
Because without that, very few people have access to either of
18
the first two components.
19
And finally, we have a basic research and human resource
20
component. Because without that, we do not maintain our
21
frontier status, nor do we have the trained personnel who will
22
not only move these frontiers forward.
But
I believe,
23
equally important an d often forgotten and are the technical people
24
who, in fact, operate these systems A make sure that they do
25
what they were designed to do.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
31
1
We have, in developing this program, set what I believe
2
are ambitious goals -- ambitious but realistic goals. As I
3
just mentioned, we seek a thousand fold improvement in useful
4
computing capability. That takes us to a trillion operations
5
per second. And & the focus here will be on developing the
6
generic technologies that will prove valuable not just in this
7
sector I but in many different industrial sectors.
8
Where appropriate, we feel it very important that the
d
9
development be performed on a cost-share A basis with industry.
10
Because we want to involve industry as deeply as we can from
11
the very beginning of this activity.
12
In software development, I think we clearly have a major
13
challenge. Because our software here in this country -- as
14
indeed I in every other country -- is now lagging behind the
15
development of hardware. And & most important, as we have
16
discussed many times before, if we are to have the rapid
17
expansion of the use of our new capacities, it is essential
18
that we develop software that is user friendly. And And of
19
course, as we both know, the high performance software of
20
today is not user friendly by the wildest stretch of the
21
imagination.
22
The National Research and Education Network is going to
23
dramatically expand and enhance the capability of the existing
inter connected computer network that L we refer to as INTERNET.
red
24
25
The overall goal here is to achieve a 100-fold increase in
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
32
1
communication speed. We want to take this up to the level of
2
gigabytes per second.
3
In addition, as you have noted many times, one of our
4
major goals here is to develop a vastly greater number of on
to
5
and off ramps on this information superhighway. There are too
6
many isolated institutions and areas in our country for a e
7
modern civilization e
8
And I hope that in the not-too-distant future our public
9
will look on this network as commonplace and as little-to-be-
That
10
feared as the telephone system. This is a goal that I think e
11
is quite within our reach.
it could have
12
If we have such a network, I think we have the
13
possibility of a catalytic effect on just about every
14
component of our society Lo on our industries, both small and
:
15
large and I appreciate the reference to small industries,
16
because I believe that perhaps there is one of the vory & major
this could be
17
potential impacts; universities and research organizations;
and
18
bus perhaps even more important the elementary and secondary
19
schools of the Nation where the real deficiencies in our
20
educational system are most apparent.
21
Finally, I think that no plan is better than its
22
execution. And & the execution of this initiative is going to
23
depend very critically on the synergy that has been developed
24
among the agencies that are participating in it. What we have
25
tried to do is to develop the plans SO that each agency does
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
33
1
what it does best I and does it in a coherent way to amplify
2
the effects of its sister agencies in this overall natural program.
3
4
constituencies l and has historical strengths. I think that / as
And / Each of the agencies, as you well know, has national
5
has been noted before, that this particular I initiative and has the
6
promise of very high payback in economic terms in social
7
terms and & as I. indicated at the outset I find it very
8
difficult to think of any other initiative that has the
9
potential of a higher payback to the American taxpayer.
10
The high-performance end of the computer market l as we we
11 all know, is relatively is small But its influence far
12
transcends it size. That is where the leading- edge
technology the leading-edge applications are developed, and and
ies and
13
1110
a
14
Federal investment in leading- edge computer technology will
15
speed the growth of the overall computer market and can
16
catalyze I I believe, investments on the part of U.S. industry.
17
But I would again come back to the matter of synergy
18
mentioned earlier. I would not want this to be thought of as
19
only a supercomputer initiative. It is much more important
20
than that. Supercomputers play a role, and a very important
21
one, but by no means the only role.
22
I think also that the initiative that we bring before you
also
23
today, has the potential to be a major contributor to meeting a
24
number of other very important national needs -- national
25
security, health, transportation, education, energy,
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
34
1
environment -- all of these are areas where the availability
and
2
of a new generation of high-performance information is handling f
3
information transfer will be essential. And the dependence, I
4
believe will, in fact, only grow.
5
If we are to realize the full potential of this
6
initiative, it is not enough that it reach its technology
e
7
goals. It is equally important I in my view, that that
8
technology be deployed by the private sector in a timely
9
fashion. And the continued development and use of Government-
10
funded, high-performance computing and communication
11
prototypes can certainly have an important, positive impact on
12
the commercialization of these technologies. And & in fact, we
13
can make available to a great many institutions in this
14
country that which cannot themselves justify through the hardware investment
15
the power that will be available an the proposed networks.
however,
l
16
This diffusion, I would submit, Mr. Chairman, is not
17
possible by Federal action alone. The administration's
18
initiative will serve the Nation best as a catalyst for
19
private action.
20
Nott
Some analysts have suggested that the initiative can spur
21
several hundred billion dollars of GNP growth. If so, it will
22
be because American companies, both small and large, have been
23
able to deploy these technologies in the production of high-
24
quality goods and services.
25
I think also that as come have predicted, that the
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
35
1
National Research and Education Network will lead to the
2
establishment -- and I hope quickly -- of a truly national,
3
high-speed network that really does connect essentially every
and
4
home A office I in the Nation. And if that happens it will be
5
because the private investments that make it possible have
6
been stimulated by the initiatives taken here.
7
Now the legislative proposals pending before the
8
Congress, I would suggest, Senator, perhaps do not fully
9
recognize the comprehensive inter I agency effort that has been
10
achieved through the years of collaboration that have led to
11
this particular activity.
NOH
12
When I testified last year T before the corresponding
now
I
13
hearing, I noted one concern and I would simply only reiterate
14
that concerno si six And that is that in a field of technology
15
that is moving as rapidly as is the case here, as you
16
illustrated with the automotive analogy, I am somewhat
17
concerned that by freezing the program in legislation we may
18
have given up some flexibility and that we may want in order to be
19
able to adjust this program,A the agency I participation, on an
20
ongoing basis.
21
I would emphasize that the FCCSET activity should still
22
be viewed as partly experimental. This is the first year that
23
we have really had it working as I had hoped it might. And
we
24
will certainly move forward next year to do a better job in
25
these areas than we have been able to this year.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
36
1
That was one concern. The other concern I have, sir,
2
reflects a fact that you mentioned. And that is e
e
3
that we have worked long and hard to bring about this total
4
integration of the agency programs. And my concern, which I
5
believe is one that you I share, is that in the
6
subsequent actions hero in of the Congress I it would be a great
7
pity of that coordination and integration were not carried
8
forward as the various pieces and players in the program
9
present their programs to your sister subcommittees.
10
And I would very much look forward to working with you,
11
sir, and ask for your assistance in making sure that we do
12
retain the coordination that is now a hallmark of this
13
activity.
14
Now I think, sir, that I woul & d, at that point I simply
15
say that I would conclude my prepared remarks. The full
16
testimony will be in the record and I would P welcome your
17
questions sirl sir
18
[The prepared statement of Dr. Bromley follows:]
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
37
1
Senator Gore: Well, thank you very much. I know my
2
colleagues will have questions also.
3
Let me begin by thanking you very much for your
4
statement. I thought it was very well put. And I, too, look
5
forward to the private investment which will be stimulated by
6
this initiative and the passage of this bill. In fact, I know
7
you are aware of the Merit proposal, with IBM and MCI and
8
MErit, announcing a massive not-for-profit venture designed
9
solely to quickly expand and proliferate the network as soon
10
as the Federal backbone network is in place.
11
I want to see a day when a school child in Tennessee can
12
come home after class and sit down, and instead of playing
13
Nintendo, use something that looks like a Nintendo apparatus
14
and plug into the Library of Congress; and read just not
15
words, but look at pictures and moving graphics presented
16
artfully and imaginatively in a way that captures and holds
17
that child's attention; responds to the child's curiosity so
18
the child can navigate through an ocean of information
19
according to what he or she wishes to explore at the moment.
20
We know how to do that. The technologies are here,
21
available, today. We ought to be empowering -- to use the
22
buzz-word of the day -- not just one or two or a few
23
individuals, but we ought to be empowering the whole country
24
to make better use of the information that is out there.
25
I also agreed with your emphasis on the level of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
38
1
cooperation that is so necessary to making this successful.
2
And, I could not agree more when you say do not freeze it. We
3
are interested in passing this and getting it moving,
4
authorizing it for the number of years necessary to complete
5
it -- sending those clear signals to the private sector that
6
it is there, it is going to be there, and it will be
7
completed, not just started, and then reevaluating every year
8
the ways to improve it and make it even better.
9
I also would like to identify with your statements about
10
the coordination of these efforts.
11
Let me start with my questions on that one. We are
12
agreed that the goals of the administration's initiative and
13
S. 272 are very, very similar, and that S. 272 is consistent
14
with what the administration is proposing.
15
Last year, as I had mentioned, the precursor, S. 1067,
16
passed the Senate unanimously. And then its passage was
17
delayed by a difference of opinion with the House, and among
18
different committees in the Senate.
19
In particular, the Department of Energy was pushing to
20
have DOE coordinate the whole program, and run the National
21
Research and Education Network. That was at-odds with the
22
administration's plan, and at-odds with the legislation.
23
I know that DOE labs like Oak Ridge and Los Alamos have a
24
critical role to play in this initiative. But I also know
25
that the best program is one that uses the strengths and
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
39
1
resources of all relevant Federal agencies.
2
Luckily, many busy people within the administration have
3
spent the last couple of years designing just such a balanced
4
program as we have. However, last year, personnel from DOE
5
and elsewhere actively lobbied against the OSTP approach and
6
bad-mouthed some of the other agencies in the program, saying
7
that they were unable to accomplish the missions that you laid
8
out for them.
9
Should we expect to see more of such bureaucratic in-
10
fighting? Or is everybody on your team on-board this year?
11
Dr. Bromley: I would devoutly hope not, Mr. Chairman. I
12
have spoken with --
13
Senator Gore: You would hope we would not see more in-
14
fighting.
15
Dr. Bromley: No more in-fighting.
16
[Laughter. ]
17
Senator Gore: Okay, I just wanted to clarify that.
18
Dr. Bromley: That is important.
19
I have spoken with the senior officials of both the
20
Department of Energy and the Department of Defense during the
21
formulation of this plan I and as we went through the FCCSET
22
process. And I have been assured by them that they are full
23
players and full participants in the program that we have
24
presented to you. We recognize that many of the agencies will
Q
25
have need for their own independent, mission-related
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
40
1
activities. But we agree completely that in order for this to
2
be a truly national program, it requires management as a
3
national program And we have identified NSF as the
l
4
appropriate agency for that. I believe that to the best of my
5
knowledge that everyone who is represented in the report that
6
you have before you is committed to moving forward with the
that report
7
program and plan laid out in there.
8
Senator Gore: Well, that is good news. And I welcome
9
that assessment. And, along with you, I devoutly hope that it
10
is the case. And I believe that it is.
11
Now, I appreciate your kind words and your statement of
12
support for the principles involved in the bill. And I fully
13
understand why any administration, given its druthers, would
14
like to have the money but no requirements on how the money is
15
spent.
16
And I take it that you -- your basic position is you like
17
the legislation. But your basic position is that you would
18
like to recommend that the Congress appropriate all of the
19
money that you have requested, but to give you no formal
20
instructions on how to spend it. Is that basically --
21
Dr. Bromley: I would phrase it somewhat differently,
22
sir.
23
[Laughter.]
24
Dr. Bromley: I would perhaps put it that we look forward
25
to working with you.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
41
1
[Laughter. ]
2
Senator Gore: Well, we look forward to working with you
3
too, Doctor.
4
Does the administration's initiative represent more than
5
just a 1-year commitment? Do you have a commitment from OMB
6
for the entire 5-year program or just for the first year?
7
Dr. Bromley: We have held, as you understand, detailed
8
discussions with not only the agencies but with OMB. And the
9
program that is before you in the 1992 budget is the one that
10
we agreed on with OMB based on a 5-year plan. But you also
11
understand, sir, that the OMB, in its normal activities,
so
so
12
focuses on the particular year And we will go back and we
13
will have to make our case again in the following year. But I
14
am convinced from our activities this year that there is
P
a
15
full recognition that what we are talking about here is an
16
important initiative It Li is a presidential initiative. And
17
that we are talking about a 5-year program to achieve what is
18
involved.
19
So, although I cannot guarantee to you anything about the
20
years beyond this, my best judgment, and that of my
21
colleagues, is that this administration is fully committed to
&
22
this as a very important and a presidential initiative.
23
Senator Gore: Of course, things change, and budgets
24
change, pressures build, lips get red, circumstances alter.
25
But formally, you have a commitment for 1 year?
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
42
1
Dr. Bromley: That is correct, sir.
2
Senator Gore: And you look forward to working with OMB
3
for the other 4 years -- just as you look forward to working
4
with us for the 5 years.
5
Dr. Bromley: I look forward to working with anyone who
6
will push this initiative forward, sir.
7
Senator Gore: All right.
8
Dr. Bromley: But let me say that I have a very
9
reasonable degree of confidence that OMB understands fully
10
that this objective is a very important one, and that it is
11
certainly the intent of everyone involved at the present time
12
to move this forward expeditiously.
13
Senator Gore: Well, I think that is a very important
14
signal to send out. I am making the point, of course, that
15
the legislation is needed, even though any administration
16
would like to have all the money for everything, each year,
17
simply appropriated and not authorized.
18
But I do not want my efforts to make that point obscure
19
the very clear signal that you are sending to the private
20
sector, to all of the agencies involved. This is going to
21
happen. This is going to happen. And everybody needs to get
22
with the program and make certain that it does happen.
23
I just have a couple more questions before deferring to
24
my colleagues. I want to explore the relationship between
25
this and other OSTP initiatives.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
43
1
Last month when the budget was released, you presented
2
two other multi-agency initiatives -- a new math and science
3
education initiative, and -- and, as it has for the past 2
4
years, the budget also included a coordinated and integrated
5
U.S. Global Change Research program involving nine different
6
agencies.
7
Is it fair to say, as I regularly do, that this program
8
will contribute greatly to these other initiatives, enhancing
9
our Nation's ability to pursue them productively. I do not
10
want to lead you on. I suspect you agree. But since these
11
other two initiatives are yours, I would like you to flesh
12
that out just a little.
13
Dr. Bromley: Well, first of all, let me say that it is
I
14
imminently e fair to make that statement. That, in fact, not
a
15
only will the high-performance computing initiatives initiative have a
not only
16
very important impact on ^ the other two ^ that you have just
17
mentioned, but, indeed, on a great many other of the
18
activities we have under way.
19
The most obvious connection has to do with the global
20
change arena where, l as you noted earlier, Bir, L we have a
21
Mission to Planet Earth that I consider very important under
22
consideration. And as was noted earlier this afternoon, once
23
we have the EOS A platform up, which will be a unique part of
will
24
that program, because a it would allow observation of individual
25
points on the earth's surface through 14 to 15 instruments
e
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
44
1
simultaneously. And & that is critically important Because
2
from those simultaneous data one can then I extract vastly more
3
than one can from just 14 or 15 individual sensors flying
4
independently. Because there is no air column correction, no
5
cloud correction, nothing of that sort.
6
However, But as was indicated, the data flow from that particular
7
remote-sensing complement alone w will send us the equivalent
8
of the Library of Congress in less than 5 days in fact. And e
9
So, unless we have the kind of speed, capacity and information
10
transfer capability that we are talking about in the high
computing and communications
11
performance ^ initiative, we simply cannot cope with the flood
12
of information that will be coming to us from the sensors that that &
13 we we have in action.
14
And perhaps even more importantly I think l is a point
15
that you touched on earlier. And that is that the human brain
16
is substantially limited with what it can do with wi in whacking through
17
great stacks of computer print-out. But on the other hand, it is
18
almost miraculously able to form hypotheses and sense patterns
19
in those same data if presented in a ^ graphical fashion. That the
20
think, is probably the largest qualitative difference that
21
high-performance computing can make in any area. Because Luss it
22
will allow us to take this flood of data and actually do
23
something with it -- make decisions, understand phenomena & that
24
would simply be beyond us otherwise) so in that area the
25
question has a very obvious answer.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
45
1
In education and mathematics, I think that the impact
2
probably is going to be greater in the long run But than even on
3
global change. But This it is not yet as obvious. ^ The fact that
4
impresses me enormously in this area is that with a single
5
fiber optic going into a classroom, every student in that
6
classroom can have self-paced, individualized instruction in
7
any subject -- repetition where repetition is necessary,
8
positive reinforcement where that has been earned and I
9
cannot think of anything that the will improve the quality of our
10
education, particularly at elementary and secondary (more school
levels, where our greatest weakness now lies / X than
the
11
12
introduction of this kind of technology into the education
13
field.
solt V in the long-run, I think the impact may be even
14
15
greater there, although it will take a little longer to bring
16-
it into place.
17
Senator Gore: I welcome that response. Just briefly, I
18
heard a presentation in one of the early hearings on this, I
19
do not know how many years ago, where someone said, and I have
20
repeated it often since then, that if one analyzes the human
21
brain in computer terms, you could say that we have a low bit
22
rate, but very high resolution.
23
The telephone company decided years ago that seven
24
numbers were the most we could remember, then they added
25
three. And yet, when we see a trillion bits of data arrayed
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
46
1
in a mosaic pattern where each has a meaning in context
2
related to all of the others, we can comprehend them all
3
almost instantaneously.
4
What the new supercomputers allow us to do is to
5
configure data in shapes and patterns over time, which enable
6
us to absorb very large quantities of it conveniently and
7
quickly. Secondly, they allow us to search through vast
8
oceans of data and instantly retrieve those particular bits
9
which are necessary to make up the particular pattern that we
10
are looking for in order to understand the problem that we are
11
trying to solve.
12
In any event, I will come to questions later. Let me
13
recognize Senator Pressler.
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
47
1
STATEMENT OF HON. LARRY PRESSLER, U.S. SENATOR FROM SOUTH
2
DAKOTA
3
Senator Pressler: Thank you very much.
4
Dr. Bromley, I am fascinated with page 26 and 27 of the
5
report which show all the agencies that have to be coordinated
6
that are involved in the program, and the agency
7
responsibilities. It is amazing to me how many of the
8
agencies -- and I am sure they are all filled with highly
9
trained scientists and highly trained people.
10
When I was in the Army in Vietnam, I was at one point on
11
detail to -- we called it ARPA then; it is now DARPA. But I
12
know the difficulty of getting highly trained people to work
13
together. You would probably have the classic job in public
14
administration or an administration in coordinating all these
15
people and getting them to work together. What is the number
16
one problem in keeping all these agencies working together on
17
this program?
18
Dr. Bromley: I would have to say, sir, that probably the
problem
19
number one program is information transfer.
20
Senator Pressler: The number one problem?
21
Dr. Bromley: Yes, because there are, as you say, a very
22
large number of people involved here, and they will work to
23
maximum effectiveness if we can be sure that everyone
24
understands what everyone else is doing, and that they are
25
really part of a coherent program. Jue & MM are devoting a very
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
48
1
large amount of our activity to trying to make sure that this
2
is the case.
3
And And I must say that I want to pay tribute here Le to the
4
people who have been involved, the people who prepared this
5
document, because they have managed what is really a
They have brought
6
remarkable feat in in bringing about a fusion of what started out
7
as separate programs in each of the agencies you see listed on
8
these pages, sir. And Ld these folk have spent many long hours
9
sitting and looking at each program and asking how does & it fit⁵
10
as part of a national program, and then adjusting wherever the
11
overlap, the duplication, the gaps were to make it actually
12
fit. And l so we already have developed a level of personal
13
communication among the members of the community involved in
all these agencies that I think l will service extraordinarily
serve vs
14
15
well in the years ahead.
16
Senator Pressler: Now, as I understand the general
17
difference -- and I am not advocating either one here
18
necessarily, because last year in one of my statements I
19
called on the administration to come forward with a plan.
20
As I understand, the basic difference is that these
21
various agencies, we would depend on you to coordinate them as
22
you saw fit; whereas, you feel the Gore-Pressler-Kasten, et
23
cetera, bill would codify too much the relationships on the
24
HPCC program regarding all these agencies. Is that a general
25
statement?
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
49
1
Dr. Bromley: I would have to modify that just a little
2
if I might, sir, because we in OSTP do not force anybody in
3
any agency to do anything but we & do when we are successful,
4
we persuade them that they have the opportunity to become part of
A
5
a much larger entity, the a national program. And so to that
6
extent, everyone here already is involved. The responsibility
7
is the agency's, but if the coordination fails, then the blame
8
is ultimately mine.
9
Senator Pressler: But basically, these are all
10
administration-appointed -- well, they are all -- the heads of
11
all these agencies are appointed by the President, usually
12
with the advice and consent of the Senate. So therefore, the
13
White House could order them to do something.
14
Dr. Bromley: In principle, yes.
15
[Laughter.]
16
Dr. Bromley: There is a famous quote that springs to
17
mind, sir. I can summon spirits from the vast deep, but will
18
they come.
19
[Laughter.]
20
Senator Pressler: Well, I know back from my days as a
21
second lieutenant in the Army listening in on meetings, to get
22
the attention of an assistant secretary was hard on some of
23
these things. If you could do that, you could accomplish
24
something, but there is great competition.
25
Dr. Bromley: May I add one word of clarification, sir?
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
50
1
One of the important things that perhaps I should have
2
emphasized is that this has been a concern within the
and
3
administration within my office. And in order III to address that
4
concern, during this past year I have restructured the FCCSET
5
committee so that its members now comprise the cabinet
6
secretaries, or deputy secretaries, and the heads of the
7
independent agencies that are responsible for all these
As arcsult,
Q
8
activities so that that ^ once the FCCSET group makes the decision,
9
that then is 9 decision that will stick from there on in
10
because it has been made by the senior officer of the agency
11
involved, and that is a key part.
FCCSET
12
Although that group is not involved in the actual
13
construction of the initiative, and it receives the initiative,
14
approves the initiative, gives the initiative its blessing
15
before it moves forward as part of the budget process.
16
Senator Pressler: I am glad to hear that, because that
17
is really key. As a point of view of public administration,
18
many of these activities require the input of highly trained
19
people, and frequently the very top people administratively in
20
these agencies are so preoccupied with the war in the Gulf or
21
with other matters that what happens, as a practical matter,
22
is that you start having these decisions made -- you have
23
assistant secretaries battling it out, so to speak, for turf.
24
And I am glad to hear that the administration has gotten
25
the very top people involved on this. I think that is key,
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
51
1
and I hope you are able to continue that level of interest.
2
And I just point that out because I am fascinated with all
3
these agencies here. This chart is an amazing Washington
4
chart, and to make all this work together with these highly
5
skilled people who perhaps do not submit to traditional
6
administrative discipline quite as much as some others, you
7
have a big job.
8
Now, let me ask a couple of questions here. Do you
9
foresee the development, at some point, of a system of user
10
fees so the supercomputer network will pay for itself
11
partially? And how would that system be structured SO that
12
user fees do not deny access to users with limited resources,
13
such as small schools and individual researchers?
14
Dr. Bromley: Wel of in the long term, sir, as I touched
15
on just briefly earlier, I look on the National Research and
16
Education Network as a pilot if you like, a precursor 1 for
17
what, as fast as we can, will become a national service that
18
is provided by the private sector as a utility service just
19
like the telephone.
20
And under those circumstances it would seem to me that we
21
would function much as we do at the present time namely that
e
22
small schools, for example, or people who wanted access to
23
this utility would make application as part of their normal
24
process of receiving support for their research activities to
25
an appropriate agency and that just as we now provide part of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
52
the
1
grants and contracts to support ^ use of telephones, copying
2
machines, and whatnot, I think L we would also be more than
3
happy to include in that list of necessary tools if you will, e
4
the charges that might be levied by those public utilities for
5
access to the computer net.
6
Senator Pressler: Now, as you know, we have EPSCOR
7
legislation to ensure that the smaller institutions are not
8
unfairly left out when Federal research grants are made. What
9
steps can be taken to ensure that the computer research and
10
development called for here in S. 272 and your proposal will
11
include small institutions? Or to put it more specifically,
12
how can we be sure an EPSCOR professor working on deep
13
drilling at the South Dakota School of Mines will be included?
14
Dr. Bromley: Well & I think perhaps the first and most
15
important answer that I can give to that is simply to tell you
16
that this is the President's wish transmitted directly to me
17
that nat if this is going to be his initiative he wants it to be
18
broadly available to institutions large and small, both in the
19
educational and in the industrial sector. And so that I look
that
can
20
on as an instruction, and we will do everything we possibly A to
21
make sure it happens. Frankly, I do not think that we will
22
achieve anything like the potential of this system unless we
23
do just what you suggest.
24
Senator Pressler: Thank you.
25
I have some additional questions for the record, Mr.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
53
1
Chairman.
2
Senator Gore: Thank you very much, Senator Pressler.
3
Senator Robb.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
54
1
STATEMENT OF HON. CHARLES S. ROBB, U.S. SENATOR FROM
2
VIRGINIA
3
Senator Robb: Thank you, Mr. Chairman. I am pleased to
4
join you and Senator Pressler and other of our colleagues on
5
this committee again as a cosponsor of this legislation, and I
6
am delighted to have a chance to be with Dr. Bromley again.
7
We have coconspired in previous incarnations on other
8
scientific projects, and it is always a pleasure to work with
9
him.
10
I did have a very brief statement that you may have
11
already included in the record. In any event, I would ask
12
unanimous consent to include it without delivering it just on
13
this particular topic.
14
Senator Gore: Without objection.
15
[The prepared statement of Senator Robb follows:]
16
[COMMITTEE INSERT]
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
55
1
Senator Robb: I just tell you, Mr. Chairman, I was
2
struck when you made mention of the fact that you would like
3
to find a way to be able to let a child at home tune in to the
4
Library of Congress. I could not help but think last night
5
when I came home, I have a child who was trying desperately to
6
do just that. She had just lost about an hour's worth of work
7
that she had done on her own PC. It had gobbled it up, and
8
she was looking for some recourse to higher authority, which I
9
was unable to provide at the moment.
10
And then when I heard Dr. Bromley say something to the
11
effect that the entire Library of Congress could be
12
transmitted in 5 days, it occurred to me that if any of these
13
children had a printer and some sort of an arrangement to
14
simply put it on automatic pilot, that we could incur costs
15
that are beyond the ability of even Congress to take care of.
16
[Laughter. ]
17
Senator Robb: I have a couple of questions that are more
18
technical in nature that I would like to submit for the
19
record, but I would like to ask Dr. Bromley just in a very
20
sense, since this is presented again in a cost-conscious
21
environment, and several of the lighter remarks and exchanges
22
took place with respect to future funding and the coincidence
23
of intent, and what have you, with respect to others here and
24
at OMB and elsewhere.
25
But I wonder if you could just talk a little bit about
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
56
1
ways that those who believe that this would be a very good
2
investment for our country, for our individual States, for our
3
industrial sector, for others, how we could attempt to provide
4
what would be traditional cost benefit analysis. How could be
5
grapple with the basic question of justifying the kinds of
6
expenditures that we are proposing here, or that are implicit
7
here, and in the long term to realize the full potential that
8
I believe and others believe this has in terms of the costs,
9
so that we could make some comparison to the costs that other
10
competing interests for the finite Government dollar might be
11
justified or evaluated.
12
Dr. Bromley: That, Senator, is an excellent question,
13
and I wish I had a better answer for you than I am going to
we
14
give you. We have back in 1989, for example do entered into a
from
15
contract with a group an Los Alamos to try to, just what you
16
have asked namely to try to come up with an economic cost
17
benefit analysis of this initiative 1 and it has been under way
18
since that time.
19
Frankly, I would have to say that I do not put too much
20
credence in the numbers that we have at the moment because e
21
the sort of range that people are talking about is that if we
22
were to implement the initiative that we have presented to
23
you, that the payback would be somewhere between $170 billion
24
and $500 billion over a period of the next 10 years. The
25
range in itself tells you a lot about how good the actual
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
57
1
calculation is; and I frankly, sir, I think that it is a little
2
premature.
3
Senator Gore: Even so, Dr. Bromley, it might be good to
4
have that on the record.
5
Dr. Bromley: It is not bad.
6
[Laughter.]
7
Senator Gore: It might be good to look at. I do not
8
want to interrupt Senator Robb's question, but those are
9
intriguing numbers. I mean, we can deal with a range.
10
[Laughter. ]
11
Senator Robb: That is right. Regrettably, very little
12
justification is frequently required around this institution.
13
If you want to believe, you do believe.
14
Dr. Bromley: Let me, then, sirs, tell you that the only
15
numbers
Life
concrete numbers that I have available to me at the
16
moment are those that came from a Gardner study that, in fact,
in the
17
was requested of back & a number of years ago/i) and and it is that
18
range $170 billion to $500 billion over 10 years. And that
19
is a very impressive payback. But I also would caution you,
since
are
20
sir, that like you being familiar with a great many cost 1
21
benefit analyses, that there are many potential pitfalls
Not
22
And I give it to you only as an indication in support of
23
my personal belief that the payoff here is probably one of the
24
best in terms of an investment viewpoint of anything that I
H
25
can conceive of us doing. We just recently, for example, had
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
58
1
the study of Professor Edwin Mansfield of the University of
2
Pennsylvania, which focused on what I. is the rate of return on
3
Federal investment in academic research and he came out with
4
a figure of 28 percent.
5
Not
Now that is a marvelous figure, because it was created
6
by a first-rank economist. And we scientists love figures of
7
that kind because our economist friends cannot argue with us
8
about them. But I would submit, sir, that if that is the
9
return across the board on Federal investment S in fundamental
10
research, then I would be prepared to wager rather heavily
11
that the return on this initiative would be higher by
12
substantial factors.
13
Senator Robb: Dr. Bromley, I cannot tell you that that
14
is encouraging news. I would only warn you that having given
15
a figure like that, you may find that in an off-set within the
16
current caps on budget, somebody will pencil in this program
17
and then spend those savings for some other program which may
18
be equally worthy.
19
[Laughter.]
20
Senator Robb: But I do appreciate it, and I think that
21
the fact that there is, at the very least, a very substantial
22
benefit in hard dollars that could saved ought to be cranked
23
into the equation sometimes. And there are occasions when we
24
spend money to get even more money back. It does not happen
25
often in government, but there are occasions when it does, and
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
59
1
in the private sector as well.
2
As I said, the other questions are really more technical,
3
and I will submit those in writing. But again, I thank you.
4
And Mr. Chairman, I thank you for not only again
5
sponsoring the legislation, but for another hearing, and
6
hopefully more understanding of the importance of this
7
particular possibility on the horizon.
8
Senator Gore: Well, Senator Robb, thank you, and thank
9
you for your early and vigorous support of the initiative in
10
the last Congress as well. And as one of the former Governors
11
who took a leading role in stimulating high-tech research and
12
development in your State, I do not know if that is the reason
13
why you were such an early cosponsor of this, but I appreciate
14
your --
15
Senator Robb: Mr. Chairman, if I could, let me just say
16
that had the economy not soured a little bit just south of the
17
Potomac, I had some proposals that we were going to use our
18
Center of Innovative Technology to house a supercomputer, but
19
got into some of the costs involved and recognized the fact
20
that the economy, in least in terms of available expendable
21
dollars, was going sour too rapidly to push it very hard.
22
But I hope when we recover, that -- I actually made a
23
presentation on this bill and the purposes of it to an
24
appropriations subcommittee that was taking a look at it, and
25
I hope at some point that we may be able to get back in as a
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
60
1
part of this overall highway that you hope to create.
2
Senator Gore: Well, thank you again for your early and
3
vigorous support. I wanted to follow up on the figures that
4
you had Dr. Bromley comment on, and I accept the caveats on
5
why models like these are not reliable. But the numbers are
6
quite significant, and they include estimates not only related
7
to GNP, but also reductions in the deficit. So your thing
8
about the offset there I know was in jest, partly in jest
9
anyway, but this --
10
Senator Robb: No, Mr. Chairman, I have learned in this
11
institution you do not jest about things like that.
12
[Laughter.]
13
Senator Gore: But this does project, as unreliable as
14
such figures are, very significant reductions in the Federal
15
budget deficit because of this. To use a more reliable way of
16
estimating its benefits, you could say that the total
17
expenditure on this program represents about 1 percent of the
18
Federal R&D budget.
19
If, therefore, the improvements in the productivity of
20
even a tiny fraction of the other 99 percent of the Federal
21
R&D budget results from this, and you know it will, then we
22
are ahead right there in terms of value saved for the
23
taxpayers before you even consider the benefits for the
24
economy.
25
It took only a tiny leap of faith to embolden those who
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
61
1
created the interstate highway program to allow them to assume
2
that it would be used when it was built. They could really
3
see that it would be used, and it was; it has been. And it
4
has vastly improved our economy. It takes an even smaller
5
leap of faith to assume that when this network is built it
6
will be used.
7
The utilization rate for the network which now exists,
8
one one-thousandth of this capacity, is growing by -- I cannot
9
remember the exact figures, but they are measured not
10
annually, but monthly. The increase is just phenomenal. And
11
so it may be hard to put reliance in specific numbers, but it
12
is easy indeed to assume that it is going to make a tremendous
13
difference for our economy.
14
I have just a few other questions, and then we will let
15
you move on. Your initiative, Dr. Bromley, places a great
16
deal of emphasis on massively parallel supercomputers.
17
Clearly massive parallelism is the only way we are going to
18
soon achieve the 1,000-fold improvements in computing power
19
needed to solve many of the grand challenges. Yet, there is
20
still a good deal of good science being done on so-called
21
conventional supercomputers.
22
And I am wondering, will the administration's initiative
23
focus solely on massively paralleled supercomputers, or will
24
funding also be provided for purchasing supercomputers that
25
are now on the market and developing new and improved software
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
62
1
for them.
2
Dr. Bromley: It is certainly, Mr. Chairman, not our
3
intention to in any way ^ suggest that the only way to go is the
4
massively parallel route. The important thing is that given a
5
problem in either science or technology, it usually turns out
6
that either the parallel or the mainframe has major advantages
7
in the solution of that particular problem,) and so & that the
8
goal here is to have both facilities on the network and
so
9
that you have essentially a transparent system.
NOH
An
10
So the individual working in Senator Pressler's
would
&
ve
11
laboratory, simply uses his work station and has access through
12
the network to whichever is most suitable to the problem that
13
happens to be addressed at the time. We in no way are
14
suggesting that we want to eliminate one of these approaches
in favor of the other. We do, however, simply note here l that
l
15
16
if we are to reach the increase in speed in the time that we
17
project here, that l the only way to do it is through the
18
scaléable massively parallel architectures.
19
Senator Gore: Well, I am a big fan of massive
20
parallelism as you know, because you and I have talked about
21
it. But I also recognize the kind of balance in the program
22
that you have just indicated with your response and how
23
important that is.
24
One other question, and it involves education. In the
25
administration's proposal, the primary justification for the
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
63
1
initiative is research and development. Frankly, I was a
2
little bit surprised that more attention was not given to the
3
educational applications of this technology. There are
4
hundreds if not thousands of ways that a national network in
5
supercomputing can help students in colleges, high schools,
6
junior highs, and even elementary schools.
7
For instance, in January I attended the annual meeting of
8
the American Library Association in Chicago, and saw a
9
demonstration of how librarians are using the NSFNET to
10
provide students with information from databases all over the
11
United States. Yet, in the administration's proposal, there
12
is almost no mention of the role that libraries will play in
13
providing information resources to other users of the NREN.
14
Was this an unintentional oversight, or does the
15
administration intend to focus almost exclusively on research?
16
And how would you personally like to see existing networks and
17
the NREN used to improve the American education system?
18
Dr. Bromley: I am in complete agreement with you,
19
Senator. And & the fact that this does not appear in this
20
particular report ssue reflects and back l on my earlier comment that this
21
is the first year -71 the first attempt built was done under immense
22
time pressure, and it was essentially building on activities
23
in which the agencies are currently involved.
24
No one in the group that developed this initiative
25
questions for a moment the tremendous importance that it will
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
64
1
have for education. And I think I can promise you that when
2
you see the report of this group next year, you will see a
3
much greater expansion of areas like are not
educations are
4
included this year because, as I say, we were working to get
5
this report to you, and there was a natural tendency to build
6
on those familiar areas in which the agencies are currently
7
involved.
8
Senator Gore: So you would not object if in our
9
authorization this year we included the educational
10
components?
11
Dr. Bromley: I believe that education is going to be a
12
very important part of this initiative, sir.
13
Senator Robb: The tough questions have got to stop.
14
Senator Gore: Yes, I know, I know. I am going to relent
15
soon, Doctor.
16
[Laughter.]
17
Senator Gore: All right. Well, I have other questions
18
like Senator Robb and Senator Pressler which I will supply for
19
the record. Is it okay if we move on now? We have had you
20
here a long time, Dr. Bromley, Dr. Wong. You have our
21
admiration for the job you are doing, and you may get tired of
22
me bragging on the good work you do in so many areas. Maybe
23
it makes up for the few --
24
Dr. Bromley: I never tire of that, Senator.
25
[Laughter.]
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
65
1
Senator Gore: Maybe it will make up for the few where we
2
disagree. But thank you so much for your leadership in this
3
area in particular. It has been a joy to work with you
4
publicly and privately, to get to know you better and, in the
5
process, help mutually to move this matter along. And we look
6
forward to continuing that working relationship.
7
Dr. Bromley: I would welcome the opportunity to work
8
with you, sir.
9
Senator Gore: Thank you very much, and thank you, Dr.
10
Wong. We will now call our panel. We have five more
11
witnesses in a panel, and then we will conclude the hearing.
12
First, Dr. Donald Langenberg, Chancellor of the
13
University of Maryland System at Adelphi, Maryland; Dr. Melvin
14
Kalos, Director of the Cornell National Supercomputer Facility
15
in Ithaca; Mr. Tracey Gray, Vice President of Marketing for
16
Government Systems with U.S. Sprint, based in Herndon,
17
Virginia; Dr. David Nagel, Vice President of Advance
18
Technology with Apple Computer, Inc.; Dr. John Wold, Executive
19
Director of the Lilly Research Laboratory, Eli Lilly & Company
20
in Indianapolis, who is accompanied by Dr. Riaz Abdulla, Head
21
of Supercomputer Applications and Molecular Design with Eli
22
Lilly.
23
Without objection, the full prepared statements of all
24
our witnesses will be included in the record. We invite you
25
to summarize what you have to present to the committee today.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
66
1
Dr. Langenberg, we will begin with you. I would like to
2
thank all of you for coming. I know that some of you have
3
traveled great distances. We really appreciate the time and
4
effort you have put into making the hearing today a useful and
5
productive one, and we will hear all of you before going to
6
questions.
7
Dr. Langenberg, please begin.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
67
1
STATEMENT OF DONALD LANGENBERG, CHANCELLOR OF THE
2
UNIVERSITY OF MARYLAND SYSTEM AT ADELPHI, MARYLAND
3
Dr -Langenberg: Thank you, Mr. Chairman. Most of my
4
biases on the issues before you stem from my service as chair
5
of a national research council panel that 2 years ago wrote a
6
report entitled Information Technology and the Conduct of
7
Research: The User's View. I will come back to that subtitle
8
in a bit.
9
I would like to make just a few points related to the
10
work of that panel and the issues before you in S. 272. The
11
panel found that there exists significant technical,
12
financial, behavioral, and infrastructural impediments to the
13
widespread use of information technology in research. And
14
though the panel's charge was confined to research, I believe
15
the same impediments exist with respect to education.
16
We made three main recommendations and a host of
17
subrecommendations. S. 272 responds to most of them, and
18
responds very well. One of the panel's principal
19
recommendations was that, and I quote, "The institutions
20
supporting the Nation's researchers led by the Federal
21
Government should develop an interconnected national
22
information technology network for use by all qualified
23
researches."
24
The National Research and Education Network responds
25
directly to the need reflected in this recommendation, and
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
68
1
also, I believe, to the very important collateral need of the
2
education sector. In my judgment, NREN, if that is the
3
correct pronunciation, will revolutionize both research and
4
education, though, of course, in a evolutionary way.
5
My third point is that when one thinks of what NREN might
6
do for education, one thinks first of the education of
7
scientists and engineers, and then perhaps of the incredible
8
potential inherent in linking NREN to every elementary school,
9
every secondary school, every public library, and every museum
10
in the country.
11
There is another educational need of utmost importance.
12
I believe that part of the challenge we face is the creation
13
of an entirely new kind of institutional infrastructure for
14
managing the new information technology, led and supported by
15
a new breed of information professionals. These may bear some
16
resemblance to librarians or to computer scientists or to
17
publishers. And whatever they might be, we need to create
18
schools for training them and institutions within which they
19
can function. And that means educational and institutional
20
innovation of a kind that S. 272 appears well designed to
21
foster
22
My fourth point is that the most important words in the
23
title of our panel report reflect the panel's most important
24
observation. And those words are "the user's view. In
25
simple terms, the panel concluded that the development of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
69
1
information technology and its applications in the conduct of
2
research -- and here I would say education, as well -- are far
3
too important to be left to the experts. The panel cautioned
4
that planning and development should be guided by users of
5
information technology, both current and prospective, not by
6
information specialists, information scientists, information
7
technologists, or local national and international
8
policymakers.
9
It may not invariably be true that the customer is always
10
right, but institutions that create technology or make policy
11
without a clear understanding and appreciation of the real
12
needs of their clients and constituents risk making serious
13
and expensive blunders. S. 272 calls for the advice of users
14
in the development of a national research and education
15
network, and I especially applaud this provision.
16
My fifth point is a very strongly held view. In my
17
preface to our panel's report I wrote, and I quote, "I share
18
with many researchers the strong belief that much of the power
19
of science whether practice by scientists, engineers, or
20
clinical researchers derives from the steadfast commitment to
21
free and unfettered communication of information and
22
knowledge," This principle has been part of the ethos of the
23
global research community for centuries, and has served it and
24
the rest humanity well.
25
If asked to distill one key insight from my service on
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
70
1
this panel I would respond with the assertion that information
2
technology is of truly enormous importance to the research
3
community, and hence to all humanity, precisely because it has
4
the potential to enhance communication of information and
5
knowledge within that community by orders of magnitude. We
6
can only now dimly perceive what the consequences of that fact
7
may be.
8
That there is a revolution occurring in the creation and
9
dissemination of information, knowledge, and ultimately,
10
understanding, is clear to me. It is also clear to me that it
11
is critically important to maintain our commitment to free and
12
unfettered communication as we explore the uses of information
13
technology in the conduct of research.
14
Now, what I asserted there about research, I would assert
15
now about education. And if I am right, and by far the most
16
profoundly important consequence of the creation of NREN will
17
not be the expedition of research or the improvement of next
18
year's balance of trade. It will be the fundamental
19
democratization of all the world's knowledge. And this means
20
placing the accumulated intellectual wealth of centuries at
21
the beck and call of every man, woman, and child.
22
What that might mean ultimately can only be guessed, but
23
let me reminisce for a moment. I grew up in a small town on
24
the Great Plains, and in that town was a Carnegie library; one
25
of hundreds Andrew Carnegie endowed across the Nation. In
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
71
1
that modest building in the modest collection of books that it
2
housed opened the world to me, and I have been grateful to
3
that Pittsburgh steel maker ever since.
4
What if I had had direct personal access to the Library
5
of Congress, the British museum, the Louvre, and the Deutsches
6
Museum all in the course of a summer afternoon in North
7
Dakota? Just imagine. Now, my point here is that there is an
8
overriding public interest in NREN and in the rest of the
9
provisions of S. 272, an interest that transcends research and
10
its industrial applications or issues of governance and the
11
timetable for commercialization. I truly believe we have an
12
opportunity here for an American achievement of truly
13
Jeffersonian proportions. Let us not blow it.
14
Finally, for my sixth point, I note with approval that
15
S. 272 identifies the National Science Foundation as the lead
16
agency for the development of NREN. The choice is wise, I
17
think. NSF has a demonstrated capacity to manage large,
18
complex, technical operations. Unlike other S&T agencies,
19
NSF's focus is not on some mission, but on its users, that is
20
to say, its client science and engineering communities.
21
And perhaps most important, alone among Federal agencies
22
NSF bears responsibility for the support of research across
23
the full spectrum of science and engineering disciplines and
24
for the training of those who perform the research and for the
25
general education in science and technology of everybody else.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
72
1
Now, Mr. Chairman, you will have gathered that I have a
2
considerable enthusiasm for S. 272; I do. I urge you and your
3
colleagues- to enact it into law.
4
Thank you.
5
[The prepared statement of Dr. Langenberg follows:]
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800.) FOR DEPO
73
1
Senator Gore: Thank you very much. We certainly
2
appreciate your forceful statement and the way you delivered
3
it.
4
Dr. Kalos, from Cornell, welcome. Swing that microphone
5
around there so we can hear you. Thank you very much for
6
coming.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
74
1
STATEMENT OF MALVIN H. KALOS, DIRECTOR, CORNELL THEORY
2
CENTER
3
Dr. Kalos: I am sitting on the end to give me ready
4
access to the computer terminal, because I am going to
5
demonstrate some.
6
Senator Gore: Very good.
7
Dr. Kalos: Mr. Chairman, it is a privilege to be invited
8
to comment on S. 272, the High Performance Computing Act of
9
1991; however, being asked to follow Dr. Bromley and others
10
makes me feel like the man who survived the Johnstown Flood.
11
The Cornell Theory Center, which is dedicated to the
12
advancement and exploitation of high-performance computing and
13
networking for science, engineering, and industrial
14
productivity, is one of the National Science Foundation
15
Supercomputer Centers. We are part of the transformation of
16
our science and engineering culture brought about by the
17
advent and adoption of high-performance computing and
18
communications in our technological society.
19
Senator Gore, Senator Pressler, Senator Robb, and the
20
other cosponsors of this bill, and the President, understand
21
the -deep and positive implications for our future.
22
Dr. Bromley has done essential work in translating these ideas
23
into effective policy. FCCSET for the first time has unified
24
the Federal approach to high-performance computing. Theirs is
25
a well-designed, well-integrated program that shows good
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
75
1
balance between the need to exploit advancing supercomputing
2
technology, the need for very high-performance networking, and
3
the need to bring these new tools to the widest possible
4
community through research and education.
5
The aim of fundamental science is to connect all our
6
knowledge in a seamless webb of quantitative understanding.
7
This is now harder to do, because we probe into more and more
8
complex phenomena that defy analysis by the mathematical tools
9
we have. Computational modeling is essential to fill this
10
need. Many areas of science involve systematic connection
11
among different phenomena at different scales of length or
12
energy. Chemistry, biology, medicine, science of materials,
13
astrophysics, are very good examples.
14
Computation is also an essential tool in experimental
15
science. The most advanced instruments, optical and radio
16
telescopes, particle accelerators, and computers themselves,
17
are studied, designed, optimized, and verified with the help
18
of computer simulation. Data collection is automated. The
19
reduction to comprehensible data sets involves enormous
20
computations in some cases. The exchange of large data sets
21
will require very heavy use of high-capacity data networks.
22
An important step in modern science, I believe, was the
23
creation by the Congress and the National Science Foundation
)
24
of the National Supercomputer Centers. That was the mark of
25
the entry by the mainstream of American research into this new
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
76
1
era of computational science. The entire scientific and
2
engineering community of the Nation has the opportunity to
3
exploit these new tools. Students and young scientists,
4
always the very heart of any important scientific change, are
5
now involved. They will carry the message to the rest of our
6
society and to the future.
7
I would like also to comment that the present program
8
includes attention to education. The NSF program, the
9
supercomputer programs include for example at Cornell the
10
Superquest program which is bringing knowledge and training of
11
supercomputing to high schools around the Country.
12
I will show some videos showing significant scientific
13
advances made possible by supercomputing, and I would like to
14
comment to Senator Pressler, in particular, that some of the
15
advances that I have in my written testimony are those that
16
come from small schools. So these centers provide this power
17
to large institutions and small, primarily research
18
institutions, primarily undergraduate institutions, and this
19
is a very important balance.
20
Another vital role of computational science is that of
21
permitting quantitative connections among different
22
disciplines. Every one of the large problems that confront
23
our society, and to whose solutions we expect science to
24
contribute, is in some sense a multidisciplinary problem.
25
Issues of the environment and medicine, to site only two,
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
77
1
involve many sciences working together; chemistry, physics,
2
engineering, fluid flow, biology, understand of materials.
3
Bringing the knowledge from these fields together to make
4
quantitative predictions about the effect of some
5
technological or regulatory proposal would be utterly
6
impossible without the use of high-performance computational
7
modeling, which is the natural language, the indispensable
8
lingua franca of quantitative multidisciplinary research. The
9
supercomputing community will soon find itself at a major
10
crossroads where the increases in performance needed for our
11
scientific mandate will demand parallel architectures.
12
To exploit these new machines, a major retooling of
13
software and algorithms will have to take place. This must be
14
started very soon if we are to make progress on the grand
15
challenge problems in the mid-1990's. The high-performance
16
computing and communications program will offer us an
17
essential opportunity to bridge the gap between today's
18
high-performance vector machines and tomorrow's highly
19
parallel systems.
20
I have emphasized how science and its application to
21
societal problems involve the national scientific community.
22
Bringing to bear this transformation of computational science
23
in the most complete and positive way requires that its
24
techniques and strategies be learned, used, and shared by the
25
widest possible group of researchers and educators. All of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
78
1
these are necessary, and the appropriate level and balance
2
among them is essential. The High Performance Computing Act
3
of 1991 is a vital step in that direction.
4
And now I will move to the screen and I am going to show
5
three videos with different scientific themes, and each has a
6
different theme to bear on the application of science to
7
industry or medicine.
8
[First videotape shown.]
9
Dr. Kalos: The first one shows the investigation, which
10
involves some important algorithmic advance by Mike Teeter,
11
who is a professor of Cornell and an engineering fellow of
12
Corning Glass.
13
We are going to see -- please start it. We are going to
14
see a model, first a simple ball and stick model, of a quartz
15
crystal. Then, the silicon and oxygen atoms will be dressed
16
in fields that represent the electron fields at various
17
densities. There are three of them, and the lowest of level
18
of electron density is shown in blue.
19
Now we see the blue level only, and we see the ramified
20
electron field that permeates the entire crystal and gives it
21
its structure. The importance of this for Corning is that
22
understanding the physics of quartz means understand the
23
physics of glass, and this was translated into making better
24
optical fibers and an important competitive advantage for
25
Corning.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
79
1
[Second videotape shown. ]
2
Dr. Kalos: The next movie is going to show something you
3
have talked about. The composition of a set of data
4
representing a sedimentary oil field in the Gulf of Mexico
5
assembled by a team of 19 organizations, 11 petroleum and 8
6
academic institutions.
7
We see a 30 by 10 mile area of the floor of the Gulf of
8
Mexico. The green zones show places where oil or gas has been
9
seen. This is a set of salt domes. They are 6 miles high
10
under the ocean. And what we are going to see -- well, now we
11
see sand and shale zones. But more important, we are going to
12
see patterns of heat-flow throughout these salt domes. And
13
the patters of heat-flow are correlated with the presence of
14
petroleum.
15
Now, this is simply the assembling for the first time of
16
disparate data from all of the partners in this. Assembling
17
it, producing this video, has presented new insights for the
18
geologists. They think they understand better than ever
19
before how to improve the recovery of petroleum from existing
20
sources worldwide. They are also going on to do serious, very
21
heavy computational modeling to try to understand in a still
22
more fundamental way the processes that are going on.
23
We are flying through the data; we are understanding what
24
is really happening.
25
Senator Gore: This is an example of what they mean when
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
80
1
they say that computational science has now joined inductive
2
reasoning and deductive reasoning as a third new branch of
3
knowledge creation.
4
Dr. Kalos: Absolutely. Yes. Incidently, these partners
5
are also far apart geographically and will need to exchange
6
these data through high-capacity networks as they work.
7
The third video has to do with the uses of ultrasound,
8
low-intensity and high-intensity ultrasound in probing the
9
human eye.
10
[Third videotape shown.]
11
Dr. Kalos: First, we will see how low-intensity
12
ultrasound is used to survey the existence of tumors in a
13
human eye. A data set is being built up, reduced to a form
14
understandable by the computer. The location of the tumor is
15
indicated. Incidently, this is the retina, which is so
16
distorted by the tumor that it is torn away from its usual
17
position in the eye. A data set is being sliced off and
18
assembled into a three-dimensional data set that the computer
19
can understand.
20
And we see in animation how this is done. We see the
21
retina there, the distorted retina. And here a
22
cómputer-usable model has been assembled. And now we see the
23
tumor in three dimensions as it rotates. Now the model is
24
used in a mathematical way to understand how the illumination
25
by high-intensity ultrasound would affect that tumor. And we
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
81
1
see the simulation of the effects of heating by a focused,
2
high-intensity ultrasound beam.
3
You see, of course, the high temperature at the center of
4
the focus. In therapy, that focus would be steered around the
5
tumor and would literally cook it into oblivion.
6
Thank you.
7
[The prepared statement of Dr. Kalos follows:]
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
82
1
Senator Gore: Boy, that was really impressive. Well, we
2
will save our questions, but thank you so much for your
3
presentation. Very impressive.
4
Our third witness on this panel is Mr. Tracey Gray, vice
5
president for marketing with the Government Systems Division
6
of US Sprint. Mr. Gray, thank you so much for joining us
7
today, and please proceed.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
83
1
STATEMENT OF TRACEY GRAY, VICE PRESIDENT OF MARKETING,
2
GOVERNMENT SYSTEMS DIVISION, US SPRINT COMMUNICATIONS COMPANY
3
LIMITED PARTNERSHIP
4
Mr. Gray: Thank you, Mr. Chairman, and Members of the
5
subcommittee. Now, these are two tough acts to follow, and I
6
do not even have my video to support me. I would like to give
7
you some comments from the perspective of the -- of a business
8
poised to take advantage and to offer to the Government and to
9
the many users of academia and industry in this country, the
10
kind of capabilities that you have designed this bill to
11
support.
12
First of all, we endorse and support this bill. I would
13
like to reference my comments, really, along the lines of how
14
we as a business in the industry see fulfilling and
15
participating in the objectives of this legislation and
16
initiative, and how we believe that we can bring some of those
17
objectives to bear.
18
I would like to depart just briefly, though, to remind
19
everybody of the power of the private sector to bring to bear
20
resources, capability, technology, and the business resources
21
and energy to meet goals of this type, providing the
22
incentives and the understanding and the perceptions are
23
properly put in place. And I think that is what this bill is
24
all about.
25
I would like to remind everybody here that just about 6
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
84
1
years ago, this country saw one of the most significant
2
economic realignments in the industry that we have ever had.
3
That was the divestiture of AT&T. US Sprint was an outgrowth
4
of that.
5
US Sprint today has a network that is just a little under
6
4 years old. It is a fully deployed, fiber-optic network that
7
transcends the Nation. It has 23,000 miles of fiber in it.
8
We support millions of residential, business, large
9
corporations and Government customers. We support 500,000 or
10
more Government customers today on that network, supporting
11
them with voice, data, video, imaging, high-speed facsimile,
12
electronic messages, packet services, and a vast array. of
13
private line services.
14
Indirectly we support hundreds of thousands of other
15
Government services through inter-operability of our networks
16
with others. We are also deploying at this time -- bear in
17
mind, this is a network slightly less than 4 years old -- we
18
are deploying at this time the capability to support sonic
19
technology.
20
What that really means is very wide band high-speed
21
network and data transmission capability within the network.
22
Bear in mind, again, this is a network slightly over 3 years
23
old. We have well over $3.5 billion invested in this network.
24
We support this bill from the standpoint that it provides
25
the seed money, the initial stimuli that we think is necessary
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
85
1
to develop the next level of applications and to bring about
2
the incentives in the private sector to make the next step,
3
which could be a quantum leap, in the deployment and the
4
investment of technology that will support the multi-gigabyte
5
transmission paths that are necessary to achieve the
6
objectives you are talking about.
7
We are very heartened to hear and see the recognition of
8
the importance of the communications link, the superhighway
9
network. We believe that the National Research Education
10
Network objectives can be achieved with this initiative, with
11
this measure.
12
I would like to bring two issues to the fore that, to
13
give us, if not some concern, some reason to watch what you
14
are doing and look at what you are doing and participate in
15
it, to ensure that these issues are dealt with. And I believe
16
Senator Pressler articulated one of those very well. And that
17
is to ensure that the funding and the development and the
18
participation is broad enough to encourage and to support the
19
users of something other than supercomputers and the users who
20
may only have to rely on multi-gigabyte networks.
21
A reason for believing this, and promoting this, is we
22
know from experience that the cost benefits and the likelihood
23
of seeing timely development, timely deployment, of the types
24
of technology you are talking about is very dependent upon a
25
broad base of users. The more fully cost can be allocated and
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. .
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
86
1
distributed along many users, the better off everybody will
2
be, and the faster and the more timely development of these
3
technologies will take in the private sector.
4
I think an example I would like to cite where we have had
5
the technology, we have had the capability for some time, and
6
yet we have not seen the full benefits and the full
7
optimization of the technology and the applications in the
8
area of integrated services digital networks, this thing
9
called ISDN.
10
Certainly the technology and the capability is there, but
11
the applications are not. And Dr. Bromley spoke eloquently
12
about the need for software and the development of basic
13
applications to drive and to take advantage of the technology
14
of that type.
15
So we encourage the planners and the architects of this
16
legislation to ensure that there is a broad participation in
17
the academic, educational, and industrial community, beyond
18
those that just rely on supercomputers. Secondly, we
19
encourage and we will do all we can through our participation
20
to ensure that the development of the network itself is a
21
development and a plan that will permit the utilization of
22
public networks to support these services and to support the
23
capabilities.
24
We do not believe that it requires a private network
25
development or application to support your objectives. We do
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
87
1
believe, and our experience tells us, that the maximum cost
2
benefits, the long-term interest of the users of a network can
3
be found with shared network applications. We have seen time
4
and again the problems that develop with private networks,
5
where you have a group of users stranded with a set
6
technology.
7
We also believe that public services and the
8
commercialization of these applications and products will
9
ensure that the Government, the need for the infusion of
10
Government money, will cease over time, will minimize over
11
time.
12
I thank you, Mr. Chairman, and members of the
13
subcommittee. That is a business perspective that differs
14
slightly from my panel members. I welcome this opportunity to
15
have spoken to you. Thank you very much.
16
[The prepared statement of Mr. Gray follows.]
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
88
1
Senator Gore: Thank you very much. Extremely valuable
2
statement. And may I say, in echoing the words of Dr.
3
Bromley, that the real-life perspectives or reality checks, if
4
you will, which we have gained from our dialogue with your
5
company and others, have been invaluable in shaping this
6
legislation. We really appreciate your statement here today
7
as well.
8
Next, Dr. David Nagel, vice president for advanced
9
technology with Apple Computers. You are invited to proceed
10
at this time. Welcome.
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
89
1
STATEMENT OF DAVID C. NAGEL, VICE PRESIDENT, ADVANCED
2
TECHNOLOGY, APPLE COMPUTER, INC.
3
Dr. Nagel: Thank you, Mr. Chairman. I am appearing
4
today not only on behalf of Apple Computer but also on the
5
behalf of the other members of the Computer Systems Policy
6
Project. We are very appreciate of the opportunity to appear
7
for the subcommittee on a favorite topic of high-performance
8
computing and networking.
9
In the fall of 1989, the 11 largest computer systems
10
companies in the U.S. formed the Computer Systems Policy
11
Project to address what we felt were some fundamental problems
12
facing our industry. It was a measure of the importance of
13
this activity that the CSPP is an association of the chief
14
executives of our companies; the CEO's are supported by the
15
chief technologist for each company, like myself, and by a
16
permanent professional staff in Washington.
17
We began our study more than a year ago with an internal
18
look at the health of our industry. We assessed technologies
19
that we believe are critical to our industry. We assessed how
20
the U.S. is doing relative to other countries and our foreign
21
competitors in those technologies, and we developed a
22
prognosis for U.S. industry performance into the future.
23
While by almost any measure, our industry is still the
24
strongest in the world, our lead appears to be diminishing
25
rapidly by almost all of the measures that we examined. In
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
90
1
1983, for example, U.S. companies held an 83 percent share in
2
the world market of computer systems, including software.
3
Between 1983 and 1989, our share of the world market declined
4
by more than 20 percent, from 83 percent to about 61 percent.
5
During the same period, Japan's share rose from 8 percent to
6
22 percent and our European colleagues' share grew from 10
7
percent to 15 percent.
8
More troubling, the computer systems industry went from a
9
significantly positive contribution to the U.S. trade balance
10
all throughout the 1980's to a position in 1990 where our
11
imports almost exactly balance our exports. While the U.S.
12
ratio of computer exports to imports moved steadily downward
13
over the last decade, Japan even more dramatically increased
14
its export-import ratio from about 2 in 1980 to more than 6 at
15
the end of the 1980's.
16
While these findings are clearly troubling to us, the
17
members of CSPP recognize that the primary burden for staying
18
competitive in the global marketplace rests squarely with our
19
own industry. So we began with an internal assessment. We
20
examined our own investment levels and competitive positions
21
in the key technologies which we think are critical to
22
success. We identified, for example, 16 critical pre-
23
competitive generic technologies, and concluded that the U.S.
24
still leads the world in half of these. And most of these are
25
software intensive.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
91
1
We also concluded that the U.S., once leading the
2
remainder, now lags the world in several critical
3
technologies, and is losing a lead in the remainder. And most
4
of these, and in contrast to the technologies for which we
5
hold a lead, the lagging technologies are mostly capital-
6
intensive ones. We also believe that, without further
7
positive action, the U.S. position will erode further in all
8
of these 16 technology areas over the next few years.
9
The computer systems industry spends 21 percent of the
10
private sector R&D, or about 10 percent of the total national
11
investment, in research and development. The investment of
12
the computer industry in 1989, more than $18 billion, is more
13
than that of any other industrial sector, and represents a 26
14
percent increase over the amount we spent in 1988, a period
15
when many other industrial sectors were reducing their R&D
16
spending.
17
In contrast to the level of investment in private
18
industry, the U.S. Government only invested about 2 percent of
19
its R&D portfolio in generic technologies related to our
20
industry.
21
Taken as a whole, we conclude that the Federal R&D budget
22
in computing is not today focused or managed in ways that are
23
needed to preserve and enhance our economic competitiveness,
24
given the rapid pace of innovation and the R&D practices of
25
other countries. In short, we believe the Federal R&D has not
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
92
1
been as helpful to the computer industry as it might be.
2
Based on our analysis and this conclusion, we have
3
outlined an initial set of technology policy recommendations.
4
These provide a strategy, we believe, for better focusing the
5
Federal R&D investment in these pre-competitive generic
6
technologies, and will help us meet very stiff international
7
competition.
8
We believe that the Government and industry must work
9
together, and jointly must take the following first steps to
10
improve the effectiveness of the spending in the U.S. First,
11
we think that there should be an improvement in the mechanisms
12
within OMB for reviewing the Federal R&D spending program. In
13
many cases, these have become so complex, it is very difficult
14
to actually figure out what is being spent.
15
Number two, we need to increase the industry input, we
16
believe, in setting Federal R&D priorities and to better
17
manage the Federal R&D budget. Number three, we think
18
industry should work with the Federal labs and with Government
19
agencies to improve, to set Federal laboratory priorities and
20
improve the return on the national investment made in these
21
labs.
22
And fourth, we look forward to working with the
23
Government in implementing high-performance computing
24
industries, including a national network capability of
25
bringing the benefits of computing to every institution,
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
93
1
household, business, and school in the Nation.
2
We have created three CEO-level working groups to address
3
our industry's participation in the Federal R&D priority
4
setting. And we are looking here at structural and legal
5
impediments of which there appear to be a variety. We are
6
increasing the degree of interaction between industry and the
7
programs in the Federal labs. And finally, we are looking at
8
ways in which we can better participate in implementing the
9
high-performance computing and communication initiatives.
10
We fully support and recommend full funding for the
11
national high-performance computing and communication
12
programs, including a National Research and Education Network.
13
We recognize and applaud the pioneering role that this
14
subcommittee and its chairman have played in recognizing the
15
importance of the development of a national information
16
infrastructure and an effective, high-performance computing
17
program.
18
We believe this efforts are critical in providing the
19
research infrastructure in maintaining our Nation's leadership
20
in basic computer and information research. The CSPP believes
21
that the high-performance computing and communication
22
initiatives will be instrumental achievement of the national
23
education in work force training goals.
24
Now, much has been written and said about the benefits of
25
high-speed networking at the institutional level, of higher
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
94
1
education levels. While we agree with and support these uses,
2
high-speed networks will allow the rate of scientific and
3
engineering progress a major grand challenge problems to
4
accelerate significantly.
5
But we also believe the benefits of high-speed networking
6
and high-performance computing should ultimately find their
7
way beyond institutions and become available to the rest of
8
us. And I am echoing this, both what my colleagues have
9
already said and what have been said by the senators.
10
I would like to briefly touch on some other benefits of
11
what we believe are a truly universal high-speed network,
12
benefits that will eventually impact a much larger number of
13
our citizens.
14
Actually, we are beginning to see the precursors of the
15
benefits of networking and distributive computing, even with
16
the overly complex low-speed network systems currently in
17
place in the U.S. Apple Computer, for example, in a project
18
called Apple Global Education, or AGE, has made our own
19
internal slow-speed electronic mail system available so that
20
school children all over the world can communicate and
21
exchange ideas.
22
For example, recently on Earth Day, 12 schools from
23
around the world collaborated and prepared and produced local
24
newspapers that featured environmental issues using this
25
network. On other networks, we have begun to see a variety of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
95
1
education applications develop. We have seen collaborations
2
between teachers in the preparation of educational materials.
3
We have seen collaborations between students conducting
4
scientific investigations. We have seen networks allow
5
students in elementary and high school to benefit from access
6
and experts in universities. And we have seen a variety of
7
on-line courses and instructional materials being prepared and
8
disseminated electronically every day.
9
Both students and teachers have access to a widening
10
range of information, databases and computing resources, all
11
remote to their physical locations. We believe these
12
applications are springing up everywhere, even though there
13
are many impediments, because our educational system is
14
discovering the value of electronic information delivery, even
15
with the very slow speed networks available today.
16
With the arrival of data exchange capabilities like those
17
that will be provided by the NREN, capabilities that will
18
allow graphics and images to be transmitted as easily as text,
19
we should see an explosion in new uses of high-speed
20
networking and education, uses which we think will
21
fundamentally transform the whole process of education.
22
Teachers and students who are in remote, rural areas --
23
in some cases, remote urban areas -- far from major libraries
24
and universities, will have access to information and
25
expertise every bit the same as their counterparts in the most
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
96
1
favored settings. Physical separation will no longer matter
2
when on-line video conferencing and other high-speed network
3
supported applications are available to every school in the
4
U.S.
5
We believe also that U.s. business will benefit from
6
high-speed networking. Apple, for example, has greatly
7
benefited from our own internal electronic mail system, as
8
have many other companies; in our case, an information
9
exchange system called Apple Link. Using Apple Link,
10
individual contributors exchange ideas and documents with one
11
another, with their managers, and with the executives of the
12
organization. From time to time, they exchange insults with
13
one another.
14
Some of the fundamental administrative activities within
15
Apple, activities that were used to generate large piles of
16
paper, now are done almost entirely electronically. The
17
advanced technology group which I head uses a video
18
conferencing network to tie together our four separate
19
physical labs in the U.S. Without this system, it is clear we
20
would all have to do a great deal more traveling and generate
21
waste along a variety of dimensions.
22
Finally, Apple uses high-performance computing to great
23
advantage both in the development of technology for products
24
and in the development of products themselves. Over the past
25
year, for example, we have used our Cray to develop advanced
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
97
1
algorithms for data compression, high-performance graphics,
2
and speech recognition, among other applications.
3
Once we prove to our satisfaction that these algorithms
4
work on our Cray, we can design special circuits, again using
5
the Cray, which makes Cray levels performance for specialized
6
applications available on our advanced personal computers at a
7
tiny fraction of the cost of a supercomputer.
8
So high-performance computing helps us both in the
9
product sense and in the technology sense.
10
In conclusion, we recognize that improving U.S.
11
technology policy is a long-term process, cannot be addressed
12
by any one organization, any single set of recommendations or
13
any given piece of legislation. Improvement of U.S.
14
technology is, nonetheless, an essential process that will
15
require the cooperative R&D investments and partnership of
16
both the private sector and the Government. We believe that
17
improving U.S. technology requires a long-term commitment and
18
a series of changes by both industry and Government over time.
19
Whether as independent CEO's or as an industry, the members of
20
CSPP are committed and will remain involved in this process.
21
Thank you very much.
22
[The prepared statement of Dr. Nagel follows.) ]
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
98
1
Senator Gore: Thank you, Dr. Nagel, and my thanks to
2
CSPP for the input. Your company, your member companies have
3
provided throughout this project -- and I now you are
4
representing the project, but please convey my personal
5
gratitude to John Sculley for his personal support and
6
encouragement of this legislation as it has gone along.
7
Dr. Nagel: I will be happy to do just that. Thank you.
8
Senator Gore: I have many questions, as I know Senator
9
Pressler does. But we have one more witness, and it is one we
10
have been looking forward to hearing: Dr. John Wold,
11
executive director with the Lilly Research Lab at Eli Lilly.
12
And you are accompanied, Dr. Wold, by Dr. Riaz Abdulla, head
13
of supercomputer applications and molecular design at Lilly.
14
Dr. Wold, please proceed.
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
99
1
STATEMENT OF JOHN S. WOLD, EXECUTIVE DIRECTOR, LILLY
2
RESEARCH LABORATORIES, ELI LILLY AND COMPANY; ACCOMPANIED BY
3
RIAZ ABDULLA, MANAGER, HIGH-PERFORMANCE COMPUTING PROGRAM, ELI
4
LILLY AND COMPANY.
5
Dr. Wold: Thank you, Mr. Chairman. Eli Lilly and
6
Company is a global corporation based in Indianapolis,
7
Indiana, that applies advancement in science to basic human
8
needs, health care, and nutrition. We compete in the
9
pharmaceutical, medical devices, diagnostic product, and
10
animal health products industries.
11
My responsibilities at Lilly include the company's high-
12
performance computing program. And with me, as you just
13
alluded to, is my colleague, Dr. Riaz Abdulla, who manages
14
this program on a day-to-day basis, and is himself a
15
practicing supercomputer user.
16
I would be pleased to have this opportunity to present my
17
company's views about the importance of a national commitment
18
to high-performance computing and to high-capacity network. I
19
am sure that this subcommittee has heard -- it will hear much
20
more in the future -- about the underlying technology required
21
to support the evolution of supercomputers and high-capacity
22
networks.
23
It is important that you share our computering
24
technologists' excitement about their visions of
25
supercomputing systems. But I think it is also important that
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
100
1
you share the visions of a research-based institution like
2
ours which have motivated us to invest in supercomputers.
3
Long-term success in the research-based pharmaceutical
4
industry depends on one factor: innovation. We must discover
5
and develop cost-effective new products that improve the
6
quality of life and offer economic benefits to patients,
7
payors and society as a whole.
8
Pharmaceutical R&D has traditionally been a high-risk,
9
complex, time-consuming, and costly enterprise. Over the past
10
half-decade, the research-based pharmaceutical industry has
11
experienced major changes. The rapid escalation of R&D costs
12
has helped precipitate major structural changes in the sector
13
of the global economy in which the U.S. is an established
14
leader.
15
An unprecedented wave of mergers, acquisitions, joint
16
ventures, has led to fewer, larger competitors. Competition
17
in the research-based pharmaceutical industry will only become
18
more challenging during the 1990's and beyond. Consequently,
19
my company, Lilly, has evaluated many opportunities to
20
reinforce its capacity to innovate, to enhance its capacity to
21
compete. Supercomputing is a case in point.
22
Our supercomputing experience was initiated through our
23
partnership with the National Center for Supercomputing
24
Applications at the University of Illinois, the NCSA. The
25
NCSA has prepared a video segment that describes our
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
101
1
involvement in their program. If we can run the video, I will
2
conclude my remarks after the video.
3
[A video was shown.]
4
Dr. Wold: Mr. Chairman, as you pointed out earlier,
5
supercomputing has created a new common language for research.
6
In recent years, scientists have developed mathematical
7
methods describing the realistic shape and motion of atoms in
8
large molecules, such as receptors or enzymes that exist in
9
the human body. These models are now emerging as important
10
tools for scientists probing new investigations into how
11
potential drug candidates would likely affect these molecular
12
targets.
13
These mathematical descriptions are based on equations
14
involving billions of numbers. Conventional computers take
15
days or weeks to perform these calculations, but
16
supercomputers can do this work in minutes or hours and permit
17
previously impossible calculations.
18
Graphic representations of the data serve as a new
19
communications medium, a new language for scientists. Teams
20
of scientists can share the same visualized image of how a
21
specific chemical agent would likely affect the receptor in
22
question. They can quickly evaluate the probable effects of
23
modifications in a chemical. They can focus the painfully
24
slow efforts required to synthesize and test new agents on
25
those compounds that appear to have the greatest potential.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
102
1
Our experience to date suggests three interrelated
2
advantages high-performance computing to our industry. These
3
systems will speed up the identification of promising drug
4
candidates. Supercomputing will enable our scientists to
5
design new drug candidates that they otherwise would not have
6
even considered.
7
These systems will foster greater collaboration among
8
scientists from various disciplines who are involved in
9
pharmaceutical research and development. Supercomputer-
10
generated graphic simulations help scientists with diverse
11
academic training to share the same vision of crucial data.
12
And thirdly, these systems will encourage truly visionary
13
exploration. Now, supercomputers are motivating our
14
scientists to ask "what if" more boldly than ever before in
15
helping them to quickly consider many possible answers to
16
their questions.
17
I want to stress that supercomputing is only a tool. But
18
it is a very powerful scientific tool, a tool that will become
19
all the more powerful with networking capabilities. A high-
20
capacity network will greatly facilitate the dynamic
21
collaboration among scientists at different locations and
22
often different institutions. The network will help us
23
optimize scarce scientific talent during a period when we are
24
almost certain to experience major shortfalls in availability
25
of highly trained scientists.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
103
1
Finally, a high-capacity network will help scientists
2
raise questions that they could never seriously ask before.
3
In conclusion, I want to stress two points. We believe that
4
supercomputers in a national, high-capacity network are
5
important to our company, to our industry, and to the medical
6
professionals and patients we serve. And we believe that
7
high-performance computing will play a crucial role in the
8
many technology-based industries and in the growth of national
9
economies that depend on these industries.
10
We strongly recommend the enactment of the High
11
Performance Computing Act of 1991 and thank you for this
12
opportunity to share our thoughts with the committee.
13
[The prepared statement of Dr. Wold follows.]
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
104
1
Senator Gore: Thank you very much.
2
Senator Pressler will have to depart for another
3
committee hearing, and I want to recognize him first.
4
Senator Pressler: Thank you very much, Mr. Chairman. I
5
am going to ask most of my questions for the record. I just
6
want to ask one question of the panel, if I may. And that is,
7
is the current fiber-optic infrastructure sufficient to handle
8
supercomputer network envisaged in S. 272 or envisaged in the
9
administration's proposal?
10
I do not think it is. But the question is how do we get
11
the fiber-optic infrastructure? How do we accomplish that?
12
Some people say we let the telephone companies, the regional
13
telephone companies into cable TV, and they will do it, so
14
they say. Others say we require the cable TV companies to
15
start laying fiber optics rather than copper.
16
And I guess a second question to that, and might be
17
addressed to Mr. Gray, is what type of user or consumer demand
18
needs to occur before private companies like yours will begin
19
to connect homes and small businesses with fiber optics to the
20
supercomputer network?
21
So, in other words, the underlying question, and some of
22
you want to think about it a little bit more, but I think that
23
is a basic question. I am working on that in the
24
communications subcommittee and some other legislation that is
25
related to this. How do we get the fiber-optic infrastructure
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
105
1
to support this?
2
Mr. Gray: I will try a first shot at some of that. It
3
certainly_will not be all inclusive. But certainly we as a
4
carrier, as well as our competitors, the more evident it
5
becomes to us and the greater the probability there is for
6
applications of the type that we are talking about here, and a
7
large user base to be established.
8
And frankly, that is what I see this initiative really
9
precipitating. It becomes a coalescing force to bring those
10
things together. We as individual members of the industry
11
cannot bring all that together; we cannot get the computer
12
industry, the users, academia, we cannot pull them together.
13
What you are proposing here does begin to coalesce those
14
forces and bring some focus, and at least provide some, some
15
perspective on our part that this could happen, as a private
16
industry. Therefore it gives us the incentive to divert and
17
to reorganize priorities to shift our investment toward these
18
kind of capabilities.
19
And within the industry, we have worked with the exchange
20
carriers; we are very dependent on them to extending the
21
capabilities of our network to their users. And more and more
22
competitive forces are at work there, because we do have
23
options other than the exchange carriers to get to our
24
customers.
25
So there are fairly powerful marketing influences that
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W. e
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
106
1
can drive this, providing there is a, the infrastructure
2
provided by the Government to fund and seed some of this, and
3
provide the stimuli to make those things happen.
4
Dr. Nagel: It may well be that the, that the political
5
regulatory and problems of that category far outweigh the
6
difficulties of getting the technical infrastructure in place.
7
I think even from the limited look we have taken at this so
8
far, the technical problems are very minimal relative to the
9
problems that we have just talked about, getting people
10
together to work on something like this.
11
Dr. Abdulla: Senator Pressler, as a user, I would like
12
to very directly state that the answer to your first question
13
is no, the existing infrastructure is not sufficient. And
14
what the proposals that we have heard today really tell us
15
about is a paradigm shift.
16
You are talking about the difference between a
17
teleprinter and a telephone, is how different we are talking
18
about. It is going to completely change the way we do things.
19
Senator Pressler: I think we have got a big job to get
20
that fiber-optic infrastructure built somehow. It is like
21
wiring the Nation, and we have to find a way to do it. The
22
big telephone companies say they can do it, but they will only
23
do it if we let them get in the cable TV business, stuff like
24
that.
25
We have to find a way, and then we have to find a way
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
107
1
that everybody has access to use, it is kind of like a gas
2
pipeline; people will have a right to have access to that
3
fiber-optics cable somehow. And if one company controls it or
4
something, or does not let other people use it, the whole --
5
it is a very difficult problem, as I see it.
6
So this whole supercomputing thing is great. But I see
7
this fiber-optic infrastructure thing as a very great problem
8
we have to solve. And I do not really have all the answers; I
9
am working on some legislation that I hope will solve part of
10
it. But any of you, if you think of any great ideas as you
11
ride the train home, tell me.
12
Senator Gore: If I could supplement this for the record,
13
let me just offer my 2 cents' worth on this. I have supported
14
the entry of telephone companies into the cable television
15
business, but that is an extremely controversial proposal
16
which may not finally pass.
17
I do believe that the measures included in this
18
legislation will result in the unleashing of forces which will
19
inevitably lead to the wiring of the Nation. In fact, the
20
fiber-optic capacity which is already in place is adequate for
21
the long distance links, provided we make available the new
22
switches, the new software, and the new algorithms, which will
23
upgrade the capacity of the existing fibers without requiring
24
the placement of new fibers in the ground or on the poles.
25
The inadequacy of the present fiber network, leaving
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
108
1
aside the software switches and algorithms, the inadequacy of
2
the fiber itself is mainly in what is called the last mile,
3
from the last switching station to the home. It is for that
4
reason that I have supported the Telco entry into cable.
5
But let us assume that that does not happen in the near
6
term. I believe very deeply that once the backbone network is
7
in place, we will witness the emergence of a new generation of
8
information services, a new generation of ways to configure
9
information to make it understandable to people, that we will
10
unleash enormous demand for access to that backbone network.
11
There will be a new set of financial incentives to
12
encourage people to provide that last mile. There will also
13
be the ongoing efforts of merit; just to name one, I know that
14
that is not a -- it is not universally participated in by
15
everybody, but it is a not-for-profit corporation, perhaps one
16
of several that will be active in rapidly expanding the reach
17
of the backbone network.
18
Just as the interstate highway system led to initiatives
19
by States and cities and even private turnpike authorities to
20
connect to the interstate highway system with new, four-lane
21
limited-access roads that were not part of the federal system,
22
as it was initially designed. So this backbone network will
23
quickly, in some cases even simultaneously, lead to the
24
completion of access links, which will themselves encourage
25
access links.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
109
1
Just as arteries and capillaries are related on down, I
2
think there will be a growing network, a growing network, with
3
lines going to more and more people.
4
What we have now is a chicken-and-egg problem. The
5
market place is not perceiving the demand for these new
6
information services because the network is not there to
7
deliver them. The market is not perceiving the demand for the
8
network to deliver them because the new services are not yet
9
there. Once that chicken-and-egg conundrum is overcome, then
10
the, we will be in a new system of supply, we will be in a new
11
reality. And the demand for these new services, I think, will
12
drive the forces that will encourage the market to complete
13
the national network.
14
Now, I believe that is a realistic vision. But just in
15
case, I also support the Telco entry into CATV, and I will
16
look forward to working on any other proposals that people
17
have to address that.
18
Dr. Langenberg: Mr. Chairman, I might be able to provide
19
some support for that view on a relatively small scale. There
20
is a generic type of community that tends to run from 20,000,
21
perhaps to 40,000 population, that contains an unusually high
22
proportion of serious users of computing facilities. It is
23
called a university.
24
And one after another over the past decade or so, I have
25
watched universities all across the country, fiber-optic
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
110
1
backbone -- provide fiber-optic backbones for themselves. It
2
is a very complex kind of drive for that effort, but it is
3
partly-demand, it is partly based on leadership of university
4
officials who can see the future coming and who want to be
5
prepared to hook into it, once it is there, as a part of
6
maintaining that institution's competitive edge. And it is
7
partly push, partly pull, but it does work.
8
Senator Gore: To use another example, the state of
9
Tennessee Public Service Commission has already embarked on a
10
very ambitious plan to provide high-performance networking to
11
virtually the entire State in anticipation of the completion
12
of the backbone network.
13
I know that there are some other states that are doing
14
the same thing. So I think that it will happen, once the
15
network is there.
16
But let me say that we have a statement for the record
17
from the U.S. General Accounting Office on supercomputing in
18
industry -- thank you very much, Senator Pressler -- and we
19
will include this for the record.
20
[The material referred to follows.]
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
111
1
Senator Gore: We also have statements for the record
2
from the American Library Association, the Computer Research
3
Association, the Association of Research Libraries, and there
4
are other associations that have also provided statements for
5
the record. Without objection, they will be included.
6
[The material referred to follows:]
7
[SUBCOMMITTEE INSERT]
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
112
1
Senator Gore: Most, if not all -- the GAO was not asked
2
to provide a statement of support or opinion one way or
3
another on this; it is a factual recounting, accounting of
4
some of the questions we addressed to them. These other
5
statements, or most, if not all, are extremely supportive of
6
the legislation.
7
Just to pick up where I left off in my last comments.
8
Dr. Nagel, you talked about making this available to the rest
9
of us, beyond the institutions. I was thinking of the supply
10
and demand forces that will be unleashed when I heard your
11
statement there, and I wanted to refer back to that.
12
Dr. Kalos, I enjoyed the videos that you showed there.
13
How many industrial partners use the Cornell facility?
14
Dr. Kalos: We have about 15 industrial partners.
15
Senator Gore: Are they concentrated in a few industries?
16
Dr. Kalos: No, they span many industries. I should also
17
mention that our major industrial partner is IBM, which is a
18
well-known manufacturer of computers of all kinds.
19
Senator Gore: I have heard of them.
20
[Laughter.]
21
Dr. Kalos: And when IBM decided to reenter the high-
22
performance computing arena, they did it in partnership with
23
Cornell University. We have been pioneers with IBM in
24
conceiving, testing, shaking down certain aspects of their
25
supercomputing, and especially their parallel computing
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
113
1
effort. And that is a partnership that will continue as IBM
2
enters the highly parallel computing arena.
3
Senator Gore: Now, I understand that Cornell runs a
4
program called Superquest to give high school students access
5
to supercomputers. How does that work?
6
Dr. Kalos: Well, we run a national competition; we
7
announce to schools around the Nation that this program is
8
available. Teams at high schools submit ideas for original
9
research that requires supercomputing for its accomplishment.
10
The proposals are evaluating by a group of independent
11
reviewers, and the winners come to Cornell. Their prize is 2
12
weeks in Ithaca --
13
[Laughter.]
14
Dr. Kalos: -- where they are provided with pizza,
15
softball, and access to high-performance computing, among
16
other essential parts of life.
17
[Laughter.]
18
Dr. Kalos: And the winners are chosen on the basis of
19
the originality of the research. Some have done videos, and
20
so on. I myself had the privilege of introducing the two
21
winners in Gainesville, Florida, and each gave a research talk
22
about research that I found interesting and original.
23
One had to do with surgical treatment for strabismus,
24
crossed eyes. And I thought it was quite original. And
25
another, in fact, was research in my own area, which is
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
114
1
stochastic simulation. And I found that extremely
2
interesting. So I was a little overwhelmed by the quality of
3
these students. And this program continues, and in fact is
4
being broadened this year with the participation of other of
5
the national centers.
6
Senator Gore: Am I wrong that relatively few U.S.
7
companies seem to be using supercomputers compared to what the
8
potential would appear to be?
9
Dr. Kalos: Compared to the potential, yes, I think that
10
is absolutely correct.
11
Senator Gore: Why is that?
12
Dr. Kalos: Well, I think that is, there are a number of
13
issues; perhaps our colleagues from Lilly could speak better
14
to this. I think first of all, the role of computational
15
modeling in science is a relatively new development. The
16
recognition by scientists is complementary to what they have
17
learned --
18
Senator Gore: Inductive and deductive reasoning.
19
Dr. Kalos: Exactly so. The idea that this is another
20
way of doing science that sheds valuable information, it is a
21
way of connecting to the knowledge they already have. This is
22
rélatively new. In addition, of course, new techniques have
23
to be learned.
24
First of all, the basic techniques of mathematical
25
representation of the problems at hand. the translation of
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
115
1
that into correct and efficient computer algorithms, the
2
realization and testing on computers of all kinds, and the
3
realization and testing especially on supercomputers.
4
So these are a number of new challenges that scientists
5
face everywhere, and I think that, as the applications grow
6
throughout the country, as our young scientists are trained in
7
computational scientists almost as a matter of course, that
8
industry will very naturally take this up.
9
Senator Gore: Dr. Wold.
10
Dr. Wold: I can say that, certainly in our case, our
11
entry into supercomputing would have been clearly impossible
12
without the national center at the University of Illinois.
13
That was our introduction. We felt it was quite a leap of
14
faith to get involved even to that level; I had not even
15
considered at that time purchasing our own supercomputer.
16
Our usage of supercomputing time at the national center,
17
as well as our own usage after we finally got our
18
supercomputer, has just in every case exceeded our
19
expectations; in fact, it exceeded our ability to plan for it.
20
So once the tool was there, the utilization just
21
increased dramatically. The key is to get that first
22
opportunity into any researcher's hands to see what can be
23
done.
24
Senator Gore: Do you work with Larry Smarr at Champaign-
25
Urbana?
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
116
1
Dr. Wold: We certainly do, yes.
2
Senator Gore: Now, geographically, you are about, what,
3
50 to 100 miles from there?
4
Dr. Wold: It is about 120 miles, yes.
5
Senator Gore: 120 miles. Do you have to go to his
6
center still? Or do you have a link?
7
Dr. Wold: We do have a link. But perhaps I could have
8
Dr. Abdulla address that, since he drove that 140 miles many
9
times.
10
Dr. Abdulla: Senator Gore, we have, and continue to have
11
an important program at the National Center for Supercomputing
12
Applications. When we started off, obviously our people had
13
to go to the center where they learned all of the different
14
kinds of technologies that Larry Smarr had over there. There
15
was a tremendous commitment to new hardware, to work stations,
16
to networks, to software, and to new algorithms.
17
We had educational programs that are ongoing, even to
18
this day. We call them FOCUS; obviously we focus on chemistry
19
and the -- it stands for Focus on Chemistry Using
20
Supercomputers. So we expose, in very tight-knit workshops,
21
the .latest in supercomputing technology. Our staff goes
22
through all these training sessions, comes back to an
23
environment which is very similar to what we have at the NCSA,
24
and continues in collaboration with consultants that we have
25
in our shop, and expand and grow.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
117
1
And you know, if you remember Jack Walton's talk in
2
Washington, D.C., when he talked about the stages of change,
3
one of the things he said was that until somebody establishes
4
a gate post, there is not going to be any diffusion.
5
So we are at that state now; we have establishes gate
6
posts. And people are saying, gosh, if it works for
7
scientists, let me try it with my problem. And then it
8
diffuses, until finally you bring about a complete revolution
9
in the way you do things. So it is a process.
10
Senator Gore: In parallel.
11
Dr. Abdulla: Yes. It is a process rather than an event.
12
And it is happening because of all of these infrastructure-
13
related items that we talked about.
14
Senator Gore: Am I wrong that Japanese companies seem to
15
be relatively more willing to explore the potential of
16
supercomputing?
17
Dr. Nagel: Well, I was going to comment. I think -- and
18
I do not necessarily have the data to support this assertion
19
-- but I think that what, one of the things that you will find
20
is that the use of high-performance computing in industry is
21
jointly, is really a function of how competitive the industry
22
is, or how competitive the people in the industry are.
23
And one of the things that we know about the Japanese is
24
that they are very, very competitive, and effectively so. So
25
I think even in the U.S., you will find those industries which
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
118
1
are the most competitive, and are, you know, frantically
2
searching for ways of getting a sustainable competitive
3
advantage, will be using advanced techniques like
4
supercomputers and high-performance computing networks and SO
5
forth, because they will give that advantage and they will,
6
you know, make that initial threshold jump to get over the
7
difficulties.
8
Senator Gore: I think the testimony of a witness from
9
Cray last year indicated that one-third of U.S. supercomputers
10
are in industry; two-thirds of Japanese supercomputers are in
11
industry. That does not clash with the impressions that you
12
all have, does it? All right.
13
Mr. Gray: Mr. Chairman, could I elaborate on that point
14
just for a moment?
15
Senator Gore: Please.
16
Mr. Gray: We happen to be well aware, through all of our
17
international negotiations and dealings, that your Japanese
18
have got very ambitious programs for upgrading their networks
19
for deploying fiber in a very ubiquitous fashion.
20
They are currently purchasing advanced technological
21
switching capability that would support multi-gigabyte
22
networks and this kind of thing, and have every evidence they
23
will be deploying this technology and is capable of, within
24
the next 2 to 3 years.
25
Senator Gore: Well, there is no question about that. We
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800). FOR DEPO
119
1
have by most estimates about an 18-month lead over the
2
Japanese in network technology. But if we choose not to
3
exploit it, we will lose it in about 18 months because they
4
are not standing still at all.
5
Dr. Nagel: I would like to make, if I might, just one
6
more comment on this business of networking, going beyond just
7
institutional support. I think one of the sometimes
8
unappreciated consequences of setting a goal that we are going
9
to move beyond sophisticated users, as Dr. Langenberg
10
mentioned you find in the universities, to people that are not
11
necessary sophisticated in the use of networking and
12
computers, is that it will force us to make them easier to
13
use, and therefore the barriers to use in industry and in
14
education and everywhere else, will be, you know, reduced
15
greatly.
16
And that is really one of the, you know, it is still a
17
fairly arcane business to use high-performance computing and
18
to use our Internet and the various range of networks that we
19
have available in this country.
20
Senator Gore: And not only with the user interface, as
21
they call it, become a lot friendlier, as they say, the cost
22
is coming down dramatically. One estimate given to the
23
subcommittee -- I do not know whether it is accurate or not
24
-- but the estimate was that a supercomputer which costs
25
between $10 million and $20 million today will almost
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
120
1
certainly with 5 years be in the $400,000 to $500,000 range.
2
If that is the case, then, and if simultaneously the ease
3
of use-improves dramatically, we will see a sudden sharpening
4
of this conflict between data processing capability, demand
5
for data processing, all on the one hand, and our ability as a
6
nation to communicate over our existing communications lines,
7
the information visualizations and packages of data that we
8
need to convey in order to communicate with each other.
9
Dr. Kalos?
10
Dr. Kalos: Mr. Chairman, I would like to amplify
11
somewhat on the issue of the contribution, in particular of
12
high-performance networking to industrial productivity. I
13
would like to call your attention to a joint program of the
14
Xerox Corporation and Cornell, Cornell Theory Center, which
15
they call the Xerox Design Research Institute.
16
That is a very broad-based program which is concerned
17
with what it takes to bring better products more quickly to
18
market. And it concerns simulation and modeling on the
19
supercomputers, but much more than that. For example, one of
20
the issues is the product history and how one learns from
21
previous products how to design better products.
22
Another issue is that Xerox, like many other companies,
23
is spread all over the country. And for people in Park, in
24
Palo Alto, to collaborate with people in Webster, near
25
Rochester, to collaborate with people in their laboratory in
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
121
1
Tarrytown, New York, and design better products that are more
2
manufacturable, that are more maintainable requires a
3
collaboration at great length.
4
And they are very concerned about their ability to
5
communicate, not only the results of supercomputer
6
calculations, but the results of many other ideas and records.
7
I consider that also important, and I believe that the present
8
bill will contribute to productivity very much.
9
Senator Gore: In what Bill Wolf again has called a "CO-
10
laboratory."
11
Dr. Kalos: Exactly.
12
Senator Gore: And I might just note for the record,
13
while we are talking about industry, that until quite
14
recently, Toyota had more supercomputers than Ford, GM and
15
Chrysler combined. That has now barely changed, but it is
16
significant.
17
Dr. Wold, could you name any products for which, that
18
Lilly has been able to develop that might not have been
19
possible in this time frame without supercomputing?
20
Dr. Wold: Well, I certainly wish I could. But our
21
business is a very long-term business in terms of research.
22
It takes 10 years after the discovery. The supercomputer
23
impacts the discovery phase of research, SO if and when we
24
have a compound, a new drug that can be linked to the
25
supercomputer, it is a number of years in the future.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
122
1
I feel, however, though, that when that day does come, we
2
probably will not remember or notice that it was discovered by
3
the supercomputer, that computational sciences will have been
4
woven into the fabric of research such that it will no longer
5
be remarkable such that we would notice that.
6
Senator Gore: I have a number of questions which I may
7
have to ask for the record because we are running out of time.
8
But Mr. Gray, one of the goals of S. 272 is to provide a
9
catalyst for development of the extensions of the network by
10
the private sector, as I was saying earlier.
11
We want the technology developed under this bill put to
12
use by commercial network providers so that every office and
13
home will have access to the information resources available
14
on the NREN. It is my view that because of the unique nature
15
of this challenge, namely, adding to fiber already in place
16
new switches and software and algorithms which will vastly
17
upgrade the capacity of the fiber.
18
And since other fiber not dedicated to this national
19
network is also there, the discovery of the new switches, et
20
cetera, will present the possibility of making them quickly
21
available for private fiber, so that the network can be very,
22
very quickly expanded. It will also serve as a sort of
23
national demonstration project, showing what is possible with
24
a national gigabyte network, and providing that a commercial
25
market exists for such services.
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
123
1
One thing that might clearly hinder development of
2
commercial, high-speed networks would be if the federal
3
government ran the NREN in a way that competed directly with
4
the private sector, and set up an unnecessary conflict.
5
S. 272 states clearly that the NREN will be phased into,
6
and I am quoting here from the bill, "will be phased into
7
commercial operation as commercial networks can meet the
8
networking needs of American researchers and educators.
9
Is that consistent with your vision of where the NREN
10
plan ought to be headed?
11
Mr. Gray: Well, that is not, not totally clear to us.
12
The objectives as articulated in the legislation are clear;
13
what is a little unclear at the moment is how we, how that
14
path will evolve and how that plan will play itself out.
15
And the point I raise is that we would like to feel
16
certain we have a role to play in our participation to ensure
17
that the initial plan on the road map that is ultimately put
18
in place and followed will assure some evolution or some, at
19
least not exclude public network opportunities to support and
20
provide those services.
21
Senator Gore: Well, I raise this question now, so as to
22
reassert for the record of this hearing, as I have in other
23
hearings, the clear intent of the sponsors of the bill and the
24
advocates of the whole project, to make that work.
25
And even, we, like you, do not yet know how to dot every
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
124
1
I or cross very T. I want you to know that that is clearly
2
our intention and that is the way it is going to happen.
3
NSF seems now to be able to let the private sector
4
provide networking services when that makes sense. They have
5
contracted with your company, with companies like MCI and ANS
6
and regional networks to run NSFNET. Do you think that NSF is
7
presently taking the right approach?
8
Mr. Gray: From what we see and what our experience has
9
been, yes.
10
Senator Gore: Okay. Well, at least we have a model to
11
work with. It may need refinement; it may need modification.
12
But we have the intent, we have the model, we have the working
13
relationship with the companies involved. So that is a good
14
place to start.
15
Suppose we did not pass S. 272. Suppose the money was
16
not appropriated. Suppose there was no Federal leadership in
17
gigabyte networking. How long would it take for the private
18
sector to start providing gigabyte networking services on its
19
own?
20
Mr. Gray: Well, I could only speculate on that. I would
21
certainly think it is in the 5-year time frame or beyond. And
22
I think the more serious issue is whether or not the process
23
would optimize the capability. I think what you are
24
suggesting there would probably create a scenario where there
25
would be test bids developed or individual applications
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
125
1
develop one by one, and directed and designed and participated
2
in by a variety of participants.
3
And while you were setting down, you would probably have
4
a piecework of applications and standards, and you would have
5
a number of potential network services out there that could
6
possibly never inter-operate with one another. I think that
7
is the bigger danger over time.
8
Senator Gore: Can you estimate what the U.S. would use
9
if we do not build a gigabyte network? Or if we do not have
10
one, that can be used in a coherent fashion?
11
Mr. Gray: It would be kind of hard to estimate something
12
like that. I would not even want to take a stab at that after
13
te kind of figures that Dr. Bromley threw around.
14
[Laughter.]
15
Senator Gore: Yeah. We could just say, maybe we could
16
just agree that it would be a lot. I really wish that we had
17
more time to explore each of these questions with follow-ups.
18
I was looking for this number earlier about the increase in
19
the amount of traffic growing in NSFNET. It has been growing
20
at between 20 percent and 30 percent per month. That is a
21
pretty rapid rate of increase.
22
Voice: Have you calculated that out to an annual
23
increase?
24
Senator Gore: About 600 percent per year. That is a
25
pretty good growth rate, I would say. And with the new
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
126
1
information services, of course, it would expand dramatically
2
beyond that.
3
Because of the pressures of time, I am going to ask each
4
one of you if you would be willing to respond to further
5
questions in writing. I appreciate that very much. We are
6
going to proceed expeditiously this year.
7
[The information referred to follows;]
8
[SUBCOMMITTEE INSERT. ]
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202) 289-2260
(800) FOR DEPO
127
1
Senator Gore: I want to express my gratitude to Chairman
2
Hollings of the full committee and Senator Danforth for their
3
support and encouragement on this whole matter. And I
4
appreciate all our witnesses here today. And I think I speak
5
for most of my colleagues on the committee in saying that --
6
most, if not all, because it passed unanimously last time --
7
in saying that we are going to move expeditiously and get this
8
done. We appreciate your help today. Thank you.
9
[Whereupon, at 5:05 p.m., the hearing was adjourned.]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
ALDERSON REPORTING COMPANY, INC.
1111 FOURTEENTH STREET, N.W.
SUITE 400
WASHINGTON, D.C. 20005
(202)289-2260
(800) FOR DEPO
Enclosure D
High Performance Computing and Communications
FY 1992 Budget
(dollars in millions)
1991
1992
1993
1894
1995
1996
DARPA
BA
183.0
232.2
263
553
399
447
OL
157.5
210.1
258
320
374
423
DOE
BA
65.0
93.0
110
138
157
OL
169
54.6
78.7
97
121
143
153
EPA
BA
1.4
5.2
5
5
5
OL
5
1.0
2.7
4
4
4
4
NASA
BA
64.0
72.4
107
134
151
OL
145
49.1
62.9
89
118
140
146
HHS/NLM
BA
13.5
17.1
17
17
17
OL
17
12.0
15.2
17
17
17
17
NIST
BA
2.1
2.9
3
3
3
OL
3
1.8
2.7
3
3
3
3
NSF
BA
169.0
213.0
262
305
354
OL
413
150.4
190.0
234
280
326
380
NOAA
BA
1.4
2.5
3
3
3
OL
3
1.2
2.1
2
2
3
3
TOTAL
BA
489.4
638.3
789
956
OL
1,087
427.7
1,202
564.6
704
865
1,009
1,129
200
OMB/ESD FAX# 3954817 OR 3953165
13:15
16/50/20
SENT BY:Xerox Telecopier 7021 ; 3-13-91 ; 5:50PM ;
2023953462-
2023951575:# 2
9/20727
GEORGE a BROWN, is., Celifornia, CHAIRMAN
ROBERT 8. WALKER, Pennsylvania
F. JAMES BENSENBAEHNER je., Wissonsin
SHERWOOD L BOEHLERT. New York
JAMES M. SCHEUER, New York
TOM LEWIS, Rorida
MARILYN LLOYD, Tennessee
DON AITTER, Pennsylvania
DAN GLICKMAN. Kansas
SID MORRISON, Washington
HAROLD L VOLKMER, Missouri
NON PACKARD. California
HOWARD WOLFS, Mianigan
RALPH M. HALL, Texas
U.S. HOUSE OF REPRESENTATIVESRECEIVE
"RAUL a. HENRY, Michigan
HARRIS W. FAWELL I
DAVE MOCURDY. Oklahoma
D. FRENCH BLAUGHTER Jan Virginis
NORMAN Y. MINITA, California
COMMITTEE ON SCIENCE, SPACE,
LAMAR SMITH. Taxas
TIM VALENTINE, North Cerolina
CONSTANCE A. MORELLA, Meryiend
ROBERT G. TOANSELLI. New Jersey
RICK BOUCHER, Virginia
AND TECHNOLOGY,
DANA ROHRABACHER, California
STEVEN M. SCHIFF, New Mexican
TERRY L BRUCE Illinois
RICHARD H. STALLINGS, leabe
91 MAR 12 P4:
TOM CAMPBELL Celifornia
OHE J. RHODES. III. Artzone
JAMES A. TRAFIGANT, Jr., Onio
SUITE 2320 RAYBURN HOUSE OFFICE BUILDING
JOE BARTON. Texas
HENRY J. NOWAK, New Yerk
DICK ZIMMER, New Jersey
CARL c. PERKINS. Kentucky
TOM MOMILLEN. Maryland
WASHINGTON, DC 20515
WAYNE T. GILCHAEST, Maryland
DAVID R. NAGLE lows
JIMMY HAYES. Louisians
(202) 225-8371
OFFICE OF THE
RADFORD BYERLY, Jr.
JERRY R COSTELLO, Illinois
JOHN TANNER, Tennessee
DIRECTOR
Chief of Staff
MICHAEL RODEMEVER
GLEN BROWDER Alabama
PETE GEREN. Texas
Chief Counsel
RAY THORNTON, Arbansas
DAVID a CLEMENT
JIM BACCHUS, Plorida
Republican Chief of Staff
TIM ROEMER, Indians
aud CRAMER. Alebama
DICK SWETT, New Harnpainire
March 11, 1991
MICHAEL J. KOPETBEL, Oregon
JOAN KELLY HORN, Missouri
BARBARA-ROSE COLLINS. Mishigan
The Honorable D. Allan Bromley
Assistant to the President for
Science and Technology
New Executive Office Building
Washington, D.C. 20506
Dear Dr. Bromley:
Thank you for your testimony last week outlining the
President's High Performance Computing and Communications
Program and your views on H.R. 656. I appreciate your
contribution to the discussion of the future of high
performance computing in this country.
Several questions remain, however, on the Administration's
position on this issue. I have attached several
additional questions for your consideration. Your
expeditious reply will greatly assist us in incorporating
your views into relevant legislation considered by the
Committee.
Thank you for your attention to this matter.
Bals Cordially, Walk
Robert S. Walker
Republican Chairman
RSW/am
Enclosure
SENT BY:Xerox Telecopier 7021 ; 3-13-91 ; 5:51PM ;
2023953462-
2023951575:# 3
COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY
U.S. HOUSE OF REPRESENTATIVES
WASHINGTON, D.C. 20515
QUESTIONS FOR THE WRITTEN RECORD
HEARING ON HIGH PERFORMANCE COMPUTING
1. Please provide the Subcommittee with both the
incremental and total run-outs for each of the agencies
participating in the President's High Performance
Computing and Communications Program (DOE, NASA, NSF,
NIST, NOAA, EPA, NIH, and DARPA) over each of the next
five years.
2. Does the President's initiative envision the government
buying and owning supercomputers, high end switches, fiber
optical cable, and other hardware? Is there any reason
why the network cannot be established commercially from
its inception? Is it necessary for the federal government
to do more than fund research and development and contract
for services from the network?
3. What specific and detailed changes would you like to