‘Superhighway’ for Data Could Speed U.S. to Top in Science, Technology

The Bush Administration plans to unveil Friday an ambitious plan to spend nearly $2 billion enhancing the nation’s technological know-how, including the creation of a high-speed data “superhighway” that would link more than 1,000 research sites around the country.

Comparable to an interstate highway system for electronic data, the network, its backers say, would move the country an important step closer to a distant dream of “wiring the nation” so that businesses and homes could receive high-definition television pictures, tap into vast electronic libraries and send video images to one another.

Establishment of the government-funded network eventually could serve as a catalyst for a massive private effort to develop the technologies needed for such high-volume transmission and to install fiber-optic links that would send information using pulses of light, according to its proponents.

‘Contributes to Leadership’

The White House “high-performance computing” initiative follows years of coordination among several government agencies and lays out various strategies for keeping the United States at the forefront of computer technology and scientific research.

A final draft of the plan argues that “a competitive domestic high-performance computer industry contributes to U.S. leadership in critical national security areas and in . . . the civilian economy.” While the five-year plan proposes federal funding for research on supercomputers and on developing software for large projects such as high-speed aircraft design and weather forecasting, its most far-reaching element is the network.

Building upon the same infrastructure of long-distance lines and fiber-optic cables that computers use to communicate now, the network would employ new transmission technologies to increase by 1,000-fold the current capacity for sending data. This would allow more users to exchange more scientific data at speeds the equivalent of 50,000 single-spaced typed pages each second.

Number of Projects Proposed

The need for an expanded network is magnified by the exploding use of supercomputers, large machines that can manipulate massive amounts of complex data at blindingly fast speeds. From cell biology to geology, from econometric modeling to weather forecasting, scientists have concocted ways to make use of mountains of data, much of which is stored in computer databases.

In addition, a number of huge scientific projects have been proposed, including mapping every gene in the human body and predicting global climate changes, that will generate masses more information. Many of these projects require linking experts from around the country so that they can share data to solve huge computational problems.

Network “communications are to technology what interstate highways are to commerce,” said Michael Connors, director of computing systems in the research division of International Business Machines Corp.

The high-volume networks also may make possible more efficient use of expensive instruments such as telescopes or medical equipment--if clear images can be transmitted instantly to medical specialists thousands of miles away, more rapid diagnosis may be available, for instance.

Experts expect that once scientific sites are connected by the proposed National Research and Education Network, commercial uses will evolve, thus giving telephone companies and other firms incentives to expand the networks and offer more services. “If it became a piece of the infrastructure . . . I’m sure there would be industrial and entertainment use,” said Robert Haber, a scientist at the National Center for Supercomputing Applications at the University of Illinois.

A similar plan has been proposed by Sen. Albert Gore (D-Tenn.), whose legislation also proposes creating a vast electronic library that could be accessed by users seeking federally gathered information. Gore proposes support to begin in fiscal 1990, while the Administration favors funding from various agencies beginning in fiscal 1991.

Both proposals, basing their estimates on a strategy report issued in 1987, call for $1.7 billion of federal funds to be spent on the supercomputing initiative, of which $400 million would go for the network.