An parallel diagnosis method for an optimal fault-tolerant network

Abstract

Recently, researchers have begun studying adaptive approaches to the diagnosable system. This paper overviews a class of optimally fault tolerant multiprocessor network architecture based on the network proposed by Pradhan and Sengupta. Because of optimal fault tolerance, the number of connections per node is precisely related to the degree of fault tolerance the network is designed to provide. Hakimi et al. assume that each unit is capable of testing every other unit using the test being conducted one at a time in such a way that the choice of the next test to be performed depends on the results of the previous tests rather than on a preselected pattern of test studied earlier. Rhee investigated the application of adaptive diagnosis approach to a D(n, t, X) system which belongs to the PMC model. In this paper, we investigate an application of the adaptive diagnosis approach to an optimally fault-tolerant multiprocessor network. The diagnosis algorithm developed here can be applied to a hypercube or a binary cube since the optimally fault-tolerant network is similar to a hypercube or a binary cube.