Design of Adptive Fault Tolerant Routing Algorithms for Hypercubes

Abstract

Interconnection networks plays an important role for the performance of modern high performance computing systems. It consists of a series of nodes and links. Each node interact with each other for communications through links. The interconnection network is a requirement of any parallel computer because parallel system shows high performance by providing reliable and quick communication over the networks. In this context, components failures have an extremely high impact because most of the routing algorithms have not been designed to tolerate faults. Since, one link and/or node failure may halt the entire computing system and stop the scientific applications running on them

Full Abstract is available in the attachment