The Defense Department is planning a major research project to create a new generation of "supercomputers," hoping to overtake Japan in a race to establish supremacy in computer technology in the 1990s, officials said yesterday.
The project, called "Strategic Computing and Survivability," but nicknamed the "Supercomputer" by Defense Department officials, is designed to compete with Japanese government and industry efforts to create a so-called "Fifth Generation" computer. Other department officials say the underlying purpose is to assure U.S. dominance over the Soviet Union in defense technology.
The new supercomputers would process information at rates thousands of times faster than existing machines and would be imbued with "artificial intelligence" software that would give them problem-solving capabilities far beyond today's computer systems.
"This is in response to the Japanese," said a high-ranking official of the Defense Advanced Research Projects Agency, which will administer the project. DARPA is tentatively seeking $50 million for fiscal 1984 and congressional hearings on the proposal are scheduled for later this month.
"The Defense Department should press this technology because no one else is pursuing it," Defense Undersecretary Richard D. DeLauer has said, "and the Japanese have strong programs in both artificial intelligence and Fifth Generation computers."
I don't think $50 million is enough," said Anthony Battista, a senior staff member on the House Armed Services Committee. Battista contends that the question of superiority in computer technology is "a problem that goes far beyond the Defense Department; it trends directly into our whole economic base."
The program was "just started up this year," said a DARPA official, and while several top computer experts are being consulted, the agency has yet to determine whether it will emphasize hardware or software development. "We are trying to resolve those issues right now," said another DARPA staffer.
It is expected that the Defense Department's supercomputer efforts would focus on new kinds of computer designs using very complex, multi-layered computer "chips" as well as "expert systems" programming that allow computers to analyze problems in much the same way that human experts do. One effort is expected to focus on using computer programs to design future computer chips.
These new technologies--and the way they can be integrated into total computer systems--have direct military applications.
"This is an enabling technology that would make advanced anti-ballistic missile systems possible," said Robert Cooper, DARPA director and assistant secretary of defense for science and technology, referring to President Reagan's recent proposal to develop missile defenses based in space.
Japan's project is sponsored by the Ministry of International Trade and Industry (MITI) which is investing $450 million over the next 10 years.
Japan expects to replace the United States as the world's leader in computers, says Edward Feigenbaum, chairman of Stanford's computer science department. "Even partially realized concepts that are superbly engineered can have great economic value, preempt the marketplace and give the Japanese the dominant position they seek," he maintains.
There are, however, industry observers who do not believe that the Japanese can wrest dominance away from the United States with a project that has yet to produce any tangible results.
Howewer, Bobby Inman, president of Microelectronics and Computer Technology Corp. (MCC), a research and development consortium of Univac, Control Data Corp. and Digital Equipment Corp., said, " The Japanese project was clearly the inspiration for the companies that put MCC together." Inman is former director of the National Security Agency.
While DARPA and MCC have similar aspirations, says Inman, "The real problem is that we'll ultimately be competing for the same talent."
This competition--plus the proprietary nature of research information in computer science and development--may create problems for both the Defense Department and the private sector in creating this proposed new generation of American supercomputer.