This project studies protocols for admitting and routing calls in distributed networks, and for tolerating failures in these networks. Its goal is to understand how to best use high-speed networking technology, such as ATM, to support parallel supercomputing, and to advance the NII (National Information Infrastructure) initiative. In the supercomputing domain, it seeks to develop network structures, routing algorithms, and mechanisms for tolerating faults that are both practical and provably effective. In the realm of the NII, it is analyzing the behavior of simple algorithms, including greedy algorithms, for admitting calls in wide-area integrated-service networks.
