"This award is recognition that the research we do here is among the best in country", says Professor of Computer Science David Nicol, the lead investigator on the grant. "NSF was impressed with the Department's depth, and our success at integrating education and research at both graduate and undergraduate levels."
The synergy in Dartmouth's research activities involves several core projects that rely on high-performance computing. Co-investigators on the project include Professor Bruce Donald, who works in computational biology and designs micro electrical-mechanical systems-- such as microscopic arms that can can position parts on assembly lines; Associate Professor David Kotz who studys parallel processing and Internet agents-- systems that harvest information for people; and Associate Professor Dan Rockmore who develops computational procedures-- based on mathematics-- for scientific processing problems. All of these computer applications use networks, remote databases, graphics and parallel processing and all share common problems-- such as effectively dividing the workload between computers that work concurrently on the same problem and avoiding network "traffic jams" when they communicate.
The connecting thread that runs through these applications is that all involve a geometric analysis of physical systems; this gives rise to common approaches to solutions. "You can look at the problem of manipulating massive amounts of data in a physical system as a sort of mathematical computation itself," says Nicol. "The work we're doing grew out of that kind of analysis, which demonstrated a way to solve the synchronization problems you encounter when you try to coordinate computers for high-performance computations."
Nicol and his Dartmouth colleagues are applying this approach to develop a model of the Internet that will allow new network management techniques to be simulated before they are deployed on the real thing. Similar approaches are being developed for solving large-scale computational problems in biology and drug design, and large-scale signal processing problems. Other projects supported by this grant include multi-media, information retrieval, and distributed robotics.