DOCUMENT RESUME IR 016 595 ED 367 315 High Performance Computing and Communications TITLE Program. Hearing before the Subcommittee on Science of the Committee on Science, Space, and Technology. U.S. House of Representatives, One Hundred Third Congress, First Session (October 26, 1993). Congress of the U.S., Washington, DC. House Committee INSTITUTION on Science, Space and Technology. REPORT NO ISBN-016-043524-2 PUB DATE 94 NOTE 110p. U.S. Government Printing Office, Superintendent of AVAILABLE FROM Documents, Congressional Sales Office, Washington, DC 20402. Legal/Legislative/Regulatory Materials (090) PUB TYPE MF01/PC05 Plus Postage. EDRS PRICE *Computer Networks; Computer System Design; Federal DESCRIPTORS Legislation; Futures (of Society); Hearings; *Industry; Information Networks; Information Systems; *Information Technology; *Technological Advancement; *Telecommunications Congress 103rd; High Performance Computing Act 1991; IDENTIFIERS *High Performance Computing and Commun Program ABSTRACT This hearing explores how the High Performance Computing and Communications Program (HPCC) relates to the technology needs of industry. Testimony and prepared statements from the following witnesses on future effects of computing and networking technologies on their companies are included: (1) F. Brett Berlin, president, Brett Berlin Associates, accompanied by David R. Audley, (2) Peter R. Bridenbaugh, managing director, Prudential Securities; executive vice president, science, engineering, environment, safety (3) Paul E. Rubbert, unit chief, and health, Aluminum Co. of America; aerodynamics research, Boeing Commercial Airplane Group; (4) W. Donald Frazer, vice president, Massively Parallel Products, Oracle Corp; and (5) Marvin G. Bloomquist, manager, information technology, Mobil Exploration and Producing Technical Center. (JLB) ********************************************************************, Reproductions supplied by EDRS are the best that can be made from the original document. \ HIGH PERFORMANCE COMPUTING AND COMMUNICATIONS PROGRAM kn J.S. DEPARTMENT OF EDUCATION ,wmi Office of Educattonat Research and Improvement Cr) EDUCATIONAL RESOURCES INFORMATION N CENTER (ERIC) n C Ims document has been reproduced as recet.tep horn the person or ofgantzatoon ono.nating . 1 C M.n01 Changes neve been made to .mprove C.) IetInOduCIIOn queIdy ;4 HEARING opintons stated tn this clOCu- PotnIs 01 vlee rnent 00 not neCeSsenly represent otitCteI OE RI posotIOn 01 Miley BEFORE THE SUBCOMMITTEE ON SCIENCE OF THE COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY U.S. HOUSE OF REPRESENTATIVES ONE HUNDRED THIRD CONGRESS FIRST SESSION OCTOBER 26, 1993 [No. 78] Printed for the use of the Committee on Science, Space, and Technology OFFICE U.S. GOVERNMENT PRINTING WASHINGTON : 1994 75-413 S Go Cl ninon Piinting (Hine Fot sale b the i. mI I lot IlIlumullS, Congrk ssional Sales Olt ice, Washington. DC 20402 ISBN 0-16-043524-2 AVAILABLE BEST COPY COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY GEORGE E, BROWN, California, Chairman JR., MARILYN LLOYD, Tennessee ROBERT S. WALKER, Pennsylvania* DAN GLICKMAN, Kansas F. JAMES SENSENBRENNER, JR., HAROLD L. VOL IMER, Missouri Wisconsin RALPH M. HALL, Texas SHERWOOD L. BOEHLERT, New York DAVE McCURDY, Oklahoma TOM LEWIS, Florida TIM VALENTINE, North Carolina HARRIS W. FAWELL, Illinois ROBERT G. TORRICELLI, New Jersey CONSTANCE A. MORELLA, Maryland RICK BOUCHER, Virginia DANA ROHRABACHER, California JAMES A. TRAFICANT, Ohio S1LVEN H. SCHIFF, New Mexico JR., JAMES A. HAYES, Louisiana JOE BARTON, Texas JOHN S. TANNER, Tennessee DICK ZIMMER, New Jersey PETE GEREN, Texas SAM JOHNSON, Texas JIM BACCHUS, Florida KEN CALVERT, California TIM ROEMER, Indiana MARTIN R. HOKE, Ohio ROBERT E. (BUD) CRAMER, Alabama NICK SMITH, Michigan JR., DICK SWETT, New Hampshire EDWARD R. ROYCE, California JAMES A. BARCIA, Michigan ROD GRAMS, Minnesota HERBERT C. KLEIN, New Jersey JOHN LINDER, Georgia ERIC FINGERHUT, Ohio PETER BUTTE, Massachusetts PAUL McHALE, Pennsylvania JENNIFER DUNN, Washington JANE HARMAN, California BILL BAKER, California DON JOHNSON, Georgia ROSCOE G. BARTLETT, Maryland SAM COPPERSMITH, Arizona ANNA G. ESHOO, California JAY INSLEE, Washington EDDIE BERNICE JOHNSON, Texas DAVID MINGE, Minnesota NATHAN DEAL, Georgia ROBERT C. SCOTT, Virginia XAVIER BECERRA, California PETER W. BARCA, Wisconsin E. Chief of Staff ROBERT PALMER, Chief Counsel MICHAEL RODEMEYER, Administrator KATHRYN HOLMES, Republican Chief of Staff D. DAVID CLEMENT, SUBCOMMITTEE ON SCIENCE RICK BOUCHER, Virginia, Chairman RALPH M. HALL, Texas SlrERWOOD L. BOEHLERT, New York TIM VALENTINE, North Carolina JOE BARTON, Texas JAMES A. BARCIA, Michigan SAM JOHNSON, Texas DON JOHNSON, Georgia NICK SMITH, Michigan ANNA G. ESHOO, California PETER BLUTE, Massachusetts EDDIE BERNICE JOHNSON, Texas DAVID MINGE, Minnesota *Ranking Republican Member. CONTENTS WITNESSES Page Dr. F. Brett Berlin, president, Brett Berlin Associates, Alexandria, VA, ac- companied by Dr. David R. Audley, managing director, Prudential Securi- ties, New York, NY; Dr. Peter R. Bridenbaugh, executive vice president, science, ezigineetzl, environment, safety and. health, Aluminum Co., of America, Pittsb , PA; Dr. Paul E. Rubbert, unit chief, aerodynamics research, Boeing Commercial Aizplane Group, Seattle, WA; Dr. W. Donald Frazer, vice president, Massively Parallel Products, Oracle Corp., Redwood Shores, CA; and Dr. Marvin G. Bloommfist, manager, information tech- nology, Mobil Exploration and Producing Technical Center, Dallas, TX 3 MGR PERFORMANCE COMPUTING AND COMMUNICATIONS PROGRAM TUESDAY, OCTOBER 26, 1993 HOUSE OF REPRESENTATIVES, COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY, SUBCOMMITTEE ON SCIENCE, Washington, DC. The subcommittee met, pursuant to notice, in room 2318, Ray- burn House Office Building, Hon. Rick Boucher (chairman of the subcommittee) presiding. Mr. BOUCHER. The subcommittee will come to order. This morning the Subcommittee on Science continues its over- sight of the High Performance Computing and Communications Program as developed by the Federal Coordinating Council for Science, Engineering, and Technology in accordance with the High Performance Computing Act of 1991. The HPCC Program is a major component of the Administration's plan for creation of an ad- vanced information infrastructure. The President's budget request for Fiscal Year 1994 was $1 billion for the program, an increase of 38 percent over the Fiscal Year 1993 funding level, indicating the importance the Administration attaches to this ongoing research initiative. In previous hearings in March of 1992 and in February of this year, the subcommittee reviewed the management, operation, and future plans for the NSFNET from which the National Research and Education Network, mandated by the 1991 legislation, will evolve. In April the subcommittee reviewed the management of the overall HPCC Program. Today we will explore how it relates to the technology needs of industry. The connection between the program and advancements in the computer and telecommunications industry is obvious, but the greatest long-term effect of the program will be in industries that make use of high performance computing and high speed networking. The potential value of these technologies to the Na- tion's economic strength was the main justification for the creation of the program and for its rapid growth. Today we have asked witnesses from several industries other than the computer and telecommunications industries to describe how the computing and networking technologies are currently used by their companies and to project how these technologies will affect their companies in the future We've asked them to tell us whether research sponsored by the program is focused in areas that will be important for the future competitiveness of their companies and we (1) 2 solicit their views concerning the allocation of research funds among the major components of the program. In addition, we would be pleased to receive from our witnesses recommendations for improvements in the implementation of the program. In particular, we're interested in the effectiveness of the National Coordination Office in communicating to the private sec- tor actions with respect to the program and goals with respect to it and activities that are sponsored by the HPCC Program. I am pleased to welcome our distinguished witnesses this morn- ing, and we will turn to them after receiving statements from other members of the subcommittee. And we'll call first on the Ranking Republican Member, the gentleman from New York, Mr. Boehlert. Mr. BOEHLERT. Thank you very much, Mr. Chairman. The underlying issue at this morning's hearing was cap:,ared well in a recent cartoon by Jeff McNelly. It merely draws a gigantic cable labeled "information superhighway" strung along the tele- phone wires until it is connected to this somewhat ramshackle house. Inside a perplexed voice wonders, "Now how do we order pizza?" I think the cartoon is right on target. The purpose of the HPCC Program is to come up with equipment and services that actually do assist people. We can't lose sight of that. This should be a pro- gram where the customer is always, or at least almost always, right. This morning we'll be looking at the program's impact on indus- try, not individuals, and we'll be examining the entire HPCC Pro- gram, not just networking. But the same question needs to be asked: will all our research result in equipment and services that serve real needs? I am not prejudging the answer to that. I've been a strong advo- cate of government involvement in high performance computing al- most since the day I arrived in Congress more than a decade ago. To make such involvement work, we must do exactly what we're doing this morning: listen to the users of the program and make sure that federal agencies are doing the sam. I look forward to hearing from this morning's witnesses. I'm sure that by the end of this hearing we'll know not only how to use the information highway to order a pizza, but how to use it to make one. And before concluding, I just want to note that Dr. Berlin is here and he has his special advisory team right behind him, two of his youngsters, including one of them I want to alert all the panelists that Rebecca is taking notes. So I look forward to this hearing and the expert witnesses. Thank you, Mr. Chairman. Mr. BOUCHER. Thank you very much, Mr. Boehlert. The gentleman from Minnesota, Mr. Minge. Mr. MINGE. I have no comment. Mr. BOUCHER. Thank you, Mr. Minge. The gentlewoman from Texas, Ms. Johnson. Ms. E.B. JOHNSON of Texas. Thank you, Mr. Chairman. I simply want to express my interest and my appreciation for your having the hearing and express my interest in knowing more about the massively parallel processing and its practical uses in 3 the industries other than the computing and telecommunications industry. So I will listen with great interest. Thank you. Mr. BOUCHER. Thank you very much. We welcome now our distinguished panel of witnesses: Dr. David Audley, the Director and Manager of Strategic Analytics and Re- search for Prudential Securities in New York; Dr. Brett Berlin, president of Brett Berlin Associates of Alexandria, Virginia, and Chairman of the IEEE Committee on Computing and Applications Infrastructure; Dr. Peter Bridenbaugh, Executive Vice President, Science, Engineering, Environment, Health and Safety for the Alu- minum Company of America; Dr. Paul Rubbert, Unit Chief, Aero- dynamics Research of the Boeing Commercial Airplane Group from Seattle, Washington; Dr. W. Donald Frazer, Vice President, Mas- sively Parallel Products of the Oracle Corporation from Redwood Shores, California; and Dr. Marvin Bloomquist, Manager for Infor- mation Technology of the Mobil Exploration and Producing Tech- nical Center in Dallas, Texas. Without objection, the prepared written statements of each of the witnesses will be made a part of the record. We would welcome your oral summaries of those statements, and in view of the num- ber of witnesses we have, we would ask that you try to keep your statements to approximately five minutes. That will give us ample time to propound questions to you. Dr. Berlin, if we may, we'd like to begin with you this morning, and we would welcome your statement. STATEMENT OF DR. F. BRETT BERLIN, PRESIDENT, BRETT BERLIN ASSOCIATES, ALEXANDRIA, VA, ACCOMPANIED BY DR. DAVID IL AUDLEY, MANAGING DIRECTOR, PRUDENTIAL SECURITIES, NEW YORK, NY; DR. PETER R. BRIDENBAUGH, EXECUTIVE VICE PRESIDENT, SCIENCE, ENGINEERING, EN- VIRONMENT, SAFETY AND HEALTH, ALUMINUM CO. OF AMERICA, PITTSBURGH, PA; DR PAUL E. RUBBERT, UNIT CHIEF, AERODYNAMICS RESEARCH, BOEING COMMERCIAL AIRPLANE GROUP, SEATTLE, WA; DR. W. DONALD FRAZER, VICE PRESIDENT, MASSIVELY PARALLEL PRODUCTS, ORA- CLE CORP., REDWOOD SHORES, CA; AND DR.. MARVIN G. BLOOMQUIST, MANAGER, INFORMATION TECHNOLOGY, MOBIL EXPLORATION AND PRODUCING TECHNICAL CEN- TER, DALLAS, TX Dr. BERLIN. Thank you, Mr. Chairman, membe, of the commit- tee. This is a particularly exciting time for us. And, particularly, what I'd like to do is start by looking back 10 years. It's very appropriate that this hearing would be at this time; it's almost exactly 10 years since the first hearing that was ever held by this committee on supercomputing. It was actually held No- vember 15, 1983, and that was significant because at that time even the scientific community was sharply divided on whether supercomputing was important. There were some scientists who came and said thatsuch as Ken Wilsonthat we would never be able to compete again if' we didn't hurry up and do something about getting supercomputing out to the scientist. But there are others who said that it wasn't 'TO 4 important, that minicomputers were fine, and that we just didn't need to move forward. So as we move into this hearing, we'd like to take a look back at where we've come from and start out by talking a little bit about what the baseline is that we established at that time. There were four reasons why the high performance computing initiative in this committee decided to move. First of all, U.S. leadership in the development and application of high performance computing was deemed fundamental and im- portant to U.S. economic competitiveness and national security. And at that time we believed that science and technology was starting to fundamentally change the way it was happening, that we were moving from a experimentally-based society to a simula- tion-based society in science and technology. And what we see here today, Mr. Chairman, is evidence that not only that has happened, but, more importantly, that those changes have begun to infiltrate the ranks of database processing, management information sys- tems, and have started to change the whole way we do decision support. Now it's important to recognize that the changes we see today are not changes because the Cold War went away or anything else. It's because everything has gone to thewe have entered the infor- mation age in a whole new way and we will never go back. And at the end of this transition, as industry by industry begins the transition to high performance computing and modeling-based deci- sionmaking, they can never go back to the old ways. That's very important. The second reason that this committee moved in 1983 was be- cause they discovered that access to U.S.-mad.. supercomputers could only be gotten by scientists if they left the United States. And, Mr. Chairman, you'll undoubtedly remember in Florida State University the rather dramatic testimony of Larry Smarr as he held up a "Scientific American" cover which had just been printed and said, "I'm proud to announce that I made this visualization. The only thing that I regret is I had to go to Germany to do it." And this committee recognized the importance of that and set in motion the National Science Foundation Center's program. Third, this committee and the scientists and engineers who came before it recognized that there was a major competition brewing as other countries started to recognize what was happening. In the last 10 years that hasthere have been dramatic changes in that. Toward the end of the eighties, for example, Japan had a rate of net new name customers that was several times the rate of new customers in the United States. They were all industrial customers. In 1985, they insisted to our Trade Ambassador that there were that Japan four they places or use only five in could supercoinputers. By 1989, they had 140-some-odd supercomputers installed. That competition has now turned the corner and we a re now starting to see the implications for that in competitiveness across the board. And, finally, we discovered that the Government, based partly on its responsibility as a customer and its potential opportunity as a large sponsor of pre-competitive R&D, could have a major impact on America's overall competitive future. 8 5 I'd like to point out what has happened since then and particu- larly focusing on the NSF. I think we need to agree that the NSF, arguably, in their Center's program spent that money and it's some of the most profitable money the NSF spent during the 1980s. Each of the centers is well-run. It serves a wide variety of user groups, and I know, Mr. Chairman, you have met all the center directors and I think you'll agree that we would characterize them as a stub- born visionary or two, and that's been very important for the suc- cess of the centers. It's also been highly leve-aged as the private sector has gone to the centers, has used them, has changed the way they do business. We're seeing some examples of that here today, and you'll hear about that later. The second aspect of that program was the modest investment in the INTERNET that the NSF began to make. Many people have forgotten that the purpose of the INTERNET was not just an ex- periment in communications; it was to connect supercomputers to each other and to the user base. And in the process of using those supercomputers we discovered a host of new applications, a host of new users, such that 40,000 networks are now connected to the INTERNET today. It's a tremendously successful and exciting pro- gram that this committee can take credit for, but also we shouldn't lose sight of. I'd like to make just a brief comment about the three areas, Mr. Chairman, that you asked me to address; first of all, the relevance to industry needs. You're going to hear, the committee is going to hear testimony by my colleagues here about very specific ways that the HPCC is meeting industry needs. Ten years ago today, we had testimony that talked about applications that could use computers a thousand times of Cray-1. We are now almost at the point where we can see on the horizon that computer coming to fruition. It's instructive to remember that those applications have not gone away. Not only have they not gone away, but we have created in the process, along the road, an entire new generation of applica- tions, and it is my belief that by the end of the nineties we will have as many applications for terriflop computing that are fun- damental to competitiveness as we have for gigaflop computing today. As part of the role that high performance computing played in industry, we have to understand that there are a couple things, factors at work here. The first is that this is not just a trend of users who want to use more computing getting access to more com- puting. What we are seeing here is fundamental change in the way science, technology, decision support, and database management is done, and that's very important. In the early eighties, SAAB as one example, was one of the first automobile manufacturers to use a Cray. They went from a VAX minicomputer directly to the use of a Cray because they had a fun- damentally new application. The National Cancer Institute, when they put their supercomputer in, went directly from a VAX mini- computer to a Cray because there was a fundamentally different application. As we look to the eighties and weto the nineties and we see the database, large databases that are starting to aggregate and 9 ci 6 the need to optimize everything about the business environment, we are seeing the ground being laid for a whole new generation of these applications. And that, indeed, is what the nineties will be. It's an age of optimization. In the eighties we sought competitive- ness in basic products. In the nineties everything that we compete on is optimized, and that includes not only airplanes, but it in- cludes tennis shoes. And if you think about it now, tennis shoe ad- vertisements emphasize high technology and optimized perform- ance as opposed to just rubber from Malaya, and it's very impor- tant. Finally, just the last couple of minutes I'd like to address briefly the role of the Federal Government in this, and I'd be happy to, in response to questionsI have a number of recommendations about the National Coordination Office, but in the interest of time I won't go into those. But the federal role in the high performance computing comes, emanates from several different aspects. The first is that the Fed- eral Government itself is a major user of the kind of things that high performance computing requires. There are many people who are walking or running around saying that, because defense needs have gotten less and the Department of Energy weapons programs have started to subside, that the federal requirement for high per- formance computing is getting less, but recall that we have every- thing from environmental cleanup to clean cars, to an entire pleth- ora of research and development requirements that we can't even begin to address yet today with the computers we have. We also have a whole new generation of the need to optimize the way we do government and the way we do government services and and databases, generations those require new of will supercomputers. And, finally, the government has several roles that are very im- portant that we often underestimate. One which I'll just pull out of my testimony is the regulatory role. In order to meet the regu- latory requirements of the nineties, the Federal Government is going to have to understand what it means to address optimized technologies. We're going to have to streamline and look at every one of the nuances of new products. Drug design is going to have to be regulated in whole new ways as we move to rational drug de- sign on the civilian side, on the commercial side. So there are at least seven or eight reasons just from functional reasons why the Federal Government has a fundamental role and interest in high performance computing, and that's before we even address the is- sues of competitiveness. Finally, I'll just close by pointing out that I think we're in one of those delightful times which I refer to as a 10-year itch. About every 10 years after a group of scientists or engineers has had to fight for large computing, they finally get a computer that in a small box equals what they fought for 10 years ago in a large box. And it happened inin the fifties, you'll recall the famous report that A.D. Little put out that said there was only need for 15 com- puters in the world. In the sixties, it frustrated Thomas Watson to know that this little guy named Seymour Cray was building a com- puter out,-in the woods that was better than anything he could do, and why would anyone want one anyway? In the seventies it was 1 0