Engineers awarded NSF grants in information technology
Jorg Liebeherr	Kevin Sullivan	Ron Williams
Kevin Skadron	Barry Johnson

By Charlotte Crystal

Five engineering faculty have won National Science Foundation research grants totaling more than $3 million for projects that seek to improve the capabilities of the computer industry -- from software to hardware to the Internet.

U.Va. computer scientists Jorg Liebeherr, Kevin Sullivan and Kevin Skadron and electrical engineers Ronald Williams and Barry W. Johnson are among the first grant recipients in the country whose work will be funded under a new strategic Information Technology Research initiative. The program, advocated by the White House and recently approved by Congress, supports fundamental research and innovative applications of information technology in science and engineering.

Liebeherr will receive more than $1.2 million, a quarter of a five-year, $5-million grant shared with John Chuang at the University of California at Berkeley, Edward Knightly at Rice University and Hui Zhang at Carnegie Mellon.

Liebeherr's project involves a revolutionary rethinking of how certain specialized services operate over the Internet. Existing methods of data transmission and storage are effective only for small private networks and cannot be expanded to accommodate use by millions of people around the world. Liebeherr's group will work to come up with new infrastructures for the Internet that will support the rapidly expanding demands for enhanced services.

Sullivan secured a grant of more than $1.3 million over three years, along with colleagues Mary Shaw at Carnegie Mellon, Barry Boehm at University of Southern California and David Notkin at University of Washington.

Sullivan's project involves transforming the way that software design is conceived and taught to prioritize the end user's needs, values and points of view. For a business, a software design that maximizes profits and minimizes costs may be the best design, while for a philanthropic foundation, a design that causes the least social trauma while having the greatest social benefit may be the best, for example.

Williams and Johnson won a three-year, $436,701 award. Their project combines education and research in an effort to improve the security of computer networks. They will identify the characteristics of effective security systems, develop a theoretical explanation for their effectiveness, and create a curriculum to teach the principles of effective network security to computer engineers.

Skadron received a two-year, $110,000 award. He is collaborating with colleagues David August and Douglas Clark, both at Princeton, on a project designed to speed up computer processing when the central processing unit (CPU) reaches a logical fork in the road. Such a fork occurs when a program encounters "if" statements, which offer different paths of resolution.

The researchers are approaching the problem simultaneously from two different directions, software and hardware. On one hand they are working to enhance the capabilities of a compiler, which translates software instructions into the 0s and 1s that computer hardware can understand. On the other hand, they want to eliminate processing bottlenecks that occur when hardware works on a complex problem while simple problems pile up behind it. These two approaches will be combined to seek new levels of cooperation between the software and hardware.

More than 1,400 research proposals were submitted to the Information Technology Research initiative. Approved were 62 large projects, averaging $1 million a year for three to five years, and another 148 smaller projects for $500,000 or less, for up to three years.