Researchers at the University of Massachusetts, Amherst, have developed a computer-based tool that, when put to use in large distributed networks, could uncover the most used, expensive and vulnerable links.
The tool would help companies and government agencies better gird their networks against natural disasters like Hurricane Katrina, and terrorist attacks, such as those on Sept. 11, 2001, by providing a way to measure network efficiencies and employ a ranking system of managed nodes and links. The information could give companies information they need to protect the most critical network components, researchers say.
"Previous measures ignored how users of networks would readjust subject to a failure and could not rank nodes and links in a reasonable way," says Anna Nagurney, John F. Smith Memorial Professor and Director of the Virtual Center for Supernetworks at UMass Amherst's Isenberg School of Management.
Nagurney and doctoral student Qiang "Patrick" Qiang have worked for the past seven months on the network-efficiency and performance-measurement project, which is funded in part by the John F. Smith Memorial Fund and in part by the National Science Foundation (NSF). The UMass Amherst team was among several groups funded by the NSF to "address issues of national security under the Management Knowledge Intensive Dynamic Systems (MKIDS) program," Nagurney says. The research supplements an exiting NSF grant on decentralized decision-making in complex networks.
The measurement tool enables a "computation of the loss in efficiency" if network components, links, or combinations of those, are damaged or destroyed in an attack, Nagurney says.
"We expect that the measure will have a wide practical use also in peacetime since it provides a quantifiable way in which to identify which network components should be best maintained based on actual usage and costs," Nagurney says. Researchers say the measure can also capture how users readjust after network disruptions. "There is a huge amount of commercial potential in this measure and its variants," she says.