Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Supercomputers: Vital Too! for the Nation's Future The Honorable Albert Gore, Jr. US. Senate Introduction of Senator Albert Gore, Jr:, by Frank Press, chairman, Na- tional Research Council: I want to welcome everyone. This is an impressive turnout from both academia and literally the most important companies in America interested in the opportunities available for supercomputers. All of us are fully aware of the exploding use of these machines the locomo- tives of the information age. They can slow time in showing the excited states of atoms and ions and chemical reactions or the behavior of quarks. They can quicken time by reshaping the earth's surface in minutes. They can help depict the unobservables the propagation of cracks in a material under a load of stress, or electrons circulating around a neutron star. But that is science, and only part of the story. There is a bewildering and growing array of applications in the devel- opment of technology and in the creation of new products and industrial processes, and that is the part of the story that this symposium addresses. In introducing Senator Albert Gore, Jr., I say with confidence that few in Congress understand better the essential role of advancing technology in this nation's future than does Senator Gore, and he has expressed that understanding by legislative leadership, by an informed and vigorous critique of federal science and technology policy, and most especially by his efforts to ensure that this nation maintains and exploits its forefront . . . position in supercomputers. As a member of the Senate Commerce Committee, he has held nu- merous hearings and introduced important legislation related to the de- velopment and implementation of supercomputers in the United States. s
6 THE HONORABLE ALBERT GORE, JR. Recently, Senator Gore held a hearing on supercomputer networks. In- troducing that hearing, Senator Gore said, "Ensuring America's world leadership in advanced computer technology may well be the most im- portant economic and technological challenge of the twenty-first century." Now he will tell us more about that. Senator Gore: Thank you very much, Prank I am delighted to be here. This is a crucial meeting, and I hope that those who might have come here with some questions in their minds about where the nation will go, and where their companies or their institutions will-go, will see this conference as a beginning point for the creation of a new national consensus about where America can go to take advantage of the revolution in supercomputing, with all it means for the future of our nation. Of course a lot of those decisions will be made on Capitol Hill, and just as the development of hardware always outpaces the development of software, so the development of software almost always outpaces the development of public policy. And true to form, we have been very slow to react. We are going to face many great challenges in the years to come- from finding shelter for 2 million homeless men, women, and children in this country to giving the next generation of Americans the best schools on earth. But I firmly believe that there may well be no more significant economic and technological challenge than pushing forward to ensure America's leadership in advanced computer technology. We have come to a turning point as a nation. That is said so frequently it almost has become a cliche, but it really is incredible to live in the present time. We have discovered the ability to destroy human life. We have found the blueprint of life itself. We have invented artificial intelligence. We are creating global environmental problems that sound like the plots of bad science fiction novels, and in all of these fields our so-called common sense is challenged because common sense is an accumulation of distilled experience, and we have moved beyond historic experience. What is happening is not only new, it is not only unprecedented, but it is also, in many cases, unimagined and is so different from what has happened in civilization up until now that we are jarred and knocked off balance and require some pause to collect ourselves and realize that this is indeed a turning point. Those of us alive today must answer a fundamental question that underlies the nuclear arms race, the greenhouse effect, the epidemics, and the starvation in the world. The underlying question is, Are we as human beings capable of rising to this unprecedented challenge? It is a question that I believe and hope will be answered affirmatively. In a sense we are confronted in field after field not just with a crisis or
SUPERCOMPUTERS: VITAL TOOL FOR THE NATION'S FUTURE 7 a problem, but with what Yogi Berra once described in his inimitable way when he said, "What we have here is an insurmountable opportunity." As we look at the possible solutions for so many of the problems that we face, we can see the emerging developments in supercomputer technology as a fabulous opportunity that must not remain insurmountable. I think that the people attending this symposium therefore have a remarkable opportunity to guide this country's future. The supercomputer is not just another useful invention. I do not know who came up with the analogy first, but I have used it many times: The supercomputer is to the Information Revolution what the steam engine was to the Industrial Revolution. We are ahead, we like to tell ourselves. We manufacture 72 percent of the supercomputers in the world. But the benefits of supercomputing do not come from the creation of the machines; they come from the use of the machines. And we are not using the machines. The companies that could be using them to open new frontiers of science and technology and competition do not have the people who can sit down and use supercomputers. Partly as a result, the companies are not buying supercomputers. And the people who do have supercomputers are not able to communicate with each other very effectively. So we do not imagine the new uses that are the most important ones, the ones that we do not understand yet. Make no mistake about it, this is a completely new field of scientific inquiry. Just as we have the two established methods of creating knowledge, inductive reasoning and deductive reasoning, so now we have computing, a totally new avenue to knowledge. We must learn to understand it and use it. Over the past year, I campaigned in every region of this country and saw the foundations of solid economic progress. But when I went to Larry Smarr's National Center For Supercomputing Applications in Urbana, Illinois, and talked with others of you in this field, I really was inspired by the potential for what can take place in America. But none of that will take place unless we solve the problems that make it an insurmountable opportunity. Frank Press referred to my hearings last month. Those hearings were part of a continuing involvement that has led me to the conviction that high-performance computing should be a top priority for the nation. We must launch an immediate assault on five fronts: 1. We must create a national fiber optic network with high capacity for linking supercomputing centers throughout the United States. 2. We must stop chiseling at the margins where funding for super- computing centers is concerned. We must take advantage of the investment that has already been made which is, after all, such a minor expenditure
8 THE HONORABLE ALBERT GORE' JR. in the scheme of things to allow this nation to take advantage of what is already in place. 3. We must put in place a special initiative to address the bottle- necks in computer software development, where an extreme shortage of specialized software for important applications is impeding our progress. 4. We must give the highest priority to educating and training young people, graduate students, and postgraduate students so that we will have the people who can help us participate In this revolution. 5. Finally, we must provide adequate funding for research and devel- opment in these related areas. Six years ago, when I introduced the idea of a national fiber optic network, I spoke with people from Corning Glass Works who were, un- derstandably, enthusiastic supporters of the concept. But I could not find widespread support for the idea. I remember as a 10-year-old child sitting in Senate hearings where my father, who was a senator at that time, introduced legislation creating the Interstate Highway System. I remember listening as the problems were confronted and discussed, and then I remember not too many years later seeing the bulldozers move the earth and seeing the drive from our farm in Carthage, Tennessee, to Washington, D.C. cut from 17 hours to 9 hours, making it a 1-day trip instead of a 2-day trip. And I remember watching the truck traffic and commerce expand exponentially. The effects on the country have been so dramatic that they have never really been cataloged or even studied intensively. It is like the effect of the telephone system on the country-it is so pervasive it is difficult to study. I began with that model and then changed it significantly as I began to explore high-performance computing. When we think and talk in the United States particularly in my profession of politics-about infrastruc- ture, we often mean highways, bridges, sewer lines, and water lines, and we need those things. But we are kidding ourselves if we think that that kind of infrastructure is the key to competing with other countries in the future. Ilying to compete on that basis is like rebuilding World War I military hardware in preparation for World War II. Infrastructure is attractive to the political system because it is a function of government. It provides a role for government to play that liberals and conservatives can both accept. The benefits, generally speaking, are available to all. All boats are lifted, to use John Kennedy's analogy to the rising tide. But in thinking about infrastructure, we have to expand our imaginations and realize that the infrastructure we most need, if our objective is to increase our nation's capacity to compete and to pursue knowledge, is going to be different from the kind of infrastructure on which we have concentrated.
SUPERCOMPUTERS: VITAL TOOL FOR THE NATION'S FUTURE 9 Can we rely on the market system to give us that kind of infrastructure? For the Interstate Highway System, we could not, yet private industry was the principal beneficiary after the government's investment was made. It was difficult to project the benefits that would generate user fees from trucks and cars paying gasoline taxes on interstate highways that did not exist, but the government made the investment and did not add a penny to the national debt, because the commerce generated by the investment was so vast as to generate user fees that have created a huge surplus in the trust fund that was established. We must commit an act of faith once again and invest in new infra- structure that we know we know will create benefits to private industry so vast as to generate user fees or some other form of compensation to more than cover the relatively small investment that would be required to create this infrastructure. Many people attending this symposium know a great deal about this network. For those who are hearing about it for the first time, let me briefly sketch the problem. Private industry, whether it is the communications industry or any other industry requiring fiber optic cable, does not yet need the kind of capacity that supercomputers need. The numbers sound big to me as a nonscientist, but 50 million bits per second is considered huge as a capacity, and the communications companies in most cases are not driving the networks much beyond that capacity. It does not make economic sense for them to do so. Supercomputers can usefully take advantage of 1 billion bits per sec- ond, or 3 billion bits per second, or 10 billion bits per second. Two years ago, with help from the House Commerce Committee, I authored and steered to passage legislation that authorized and kicked off a wide-ranging study of high-performance computing networks by the Office of Science and Technology Policy. There are already more than 100 major networks in the country, including the NSFNet, but coordination among them is limited. The study found, as I expected, that these superhighways for informa- tion are now more like left-turn lanes at rush hour. They have low capacity. They are overloaded. They are unable to keep pace with demand. This study also warned that, while the United States continues to develop the best supercomputer technologies, we have been less than successful in ap- plying them to address our needs. Once again we are in danger of inventing a technology only to watch other nations apply the technology. As John Young has said, Silicon Valley is not very different from Detroit if the latest trade figures for electronics are examined. The Japanese are not that far behind, because although we invented electronic devices and created them first, we have not been using them.
10 THE HONORABLE ALBERT GORE, JR. John Connolly at the University of Kentucky's Center for Computa- tional Sciences said at the hearing that computer users will be able to send high-density applications such as high-quality pictures and graphics through supercomputer networks, but that demand for capacity far exceeds supply. He said that the nation may soon find itself in a "graphic jam." A few years ago I noticed that the Japanese entity for targeting key technologies for accelerated funding and attention had produced a list of 10 or 12 top-priority projects. One of them was a 10-gigabaud a 10- billion-bits-per-second fiber optic network. What are we doing? We need to build on the advantages that we have in this field before it is too late. It is getting to the point that it is almost too late. But it is not yet too late. That is the good news. We have the capacity to move quickly. But I want to argue to the National Academy of Sciences, to the commercial entities represented in this symposium, and to scientists and researchers from fields other than computer science that the creation of this national high-volume fiber optic network, a superhighway for information linking supercomputing centers, ought to be the number-one science priority for the United States of America. It is not just a computer science project. We all know from studying the history of science about the intimate link between communications capacity and scientific advance. Why did scientific discovery explode after the invention of the printing press? Why do so many important advances occur almost simultaneously at different points around the world? It is, of course, because communication makes it possible to assemble all of the pieces of a mosaic that then becomes apparent to many people at the same time. Now the communications technology of tomorrow, the computing capacity inherent in supercomputers, is virtually impossible for us to use because we cannot link supercomputer centers. Think for a moment with me what it would be like if clusters of researchers in universities and commercial enterprises all over America could share the capacity of those machines and the infrastructure existing at supercomputing centers, and then communicate on a regular basis with their counterparts all over the United States. The word synergy is inadequate to describe the advances that would occur. Imagine for a moment what it would be like if state governments competing one with another built interchanges to connect to this network, helping to create clusters of small businesses entering the information industry and able to download their products for distribution to a hungry market connected to the network at other points throughout the United States. There are few things we could do that would contribute more to this nation's ability to compete effectively in the future. Many questions have
SUPERCOMPUTERS: VITAL TOOL FOR THE NATION'S Fl¢URE 11 to be answered before the project I propose can become a reality. The questions are being attacked right now, but the commitment must be there. I believe that those who are part of this field should turn their attention to this network as a high priority. I believe that access to the incredible capabilities of supercomputers on a broad scale would of itself pull more bright young people into the field and into programs where they could acquire the skills that would enable them to then go into industry and teach their employers how to revolutionize the particular businesses in which they were working. Another vital concern that we must address is the bottleneck in soft- ware development. I have introduced legislation that is focused particularly on educational software and again is designed to address a particular area that market incentives are not solving. There are specialized applications for software that have a relatively small market at this point, and the money to pay for the software is really not there, but the payoff from that software is incredibly large. We ought to understand that, we ought to fill that gap, and we ought to create incentives where none now exist to produce that software. I have called for the establishment of a public-private corporation that will evaluate particularly high-priorit,r projects and then provide seed money that has to be leveraged by private investment with a large multiple to the public investment. But the private investment will be pulled in by the imprimatur or seal of approval that comes from the selection of a particular software project that is greatly needed and has high priority. I think the basic idea has worked in the past, and I think that it can work in the future. The inattention to education is an old story. I mentioned John Young earlier; he and his fabulous commission have received virtually no atten- tion. But their report (Picking Up the Pace: The Commercial Challenge to American Innovation, Council on Competitiveness, Washington, D.C., 1988) is epochal, and it emphasizes education-pre-18 education, education in computer science and technology, and education in the traditional basic sciences. We must face up to the problems in the educational system. We must also address the other problem that the Young Commission focused on. That is, how do we fill the gap between research and appli- cations? Senator Fritz Hollings, chairman of the Commerce Committee, has come up with some imaginative proposals, as have others. I intend to spend much of the next 2 years focusing on that particular problem, and I welcome input from many here who have thought a great deal about it. And then finally, we have to have adequate funding for the supercom- puter centers and for research and development across the board. I believe we need leadership. I believe we need vision. I believe we need commitment. Because we know what can be done. We know what the payoff can be. In a high-speed, high-stakes competition with the Japanese,
12 THE HONORABLE ALBERT GORE, JR. mild words of encouragement are simply not good enough. You know that, and I want you to feel one day soon that your government also knows that. Each of you in your field is doing America a favor by pointing the way to the future. The Japanese have proved what a nation can accomplish with powerful ideas and determination, and I believe it is America's turn to do the same. After all, we were the ones who showed them how, and it is up to us to renew the American spirit. I believe we are up to the task.