Conditional Diagnosability of (n, k)-Star Networks Under the Comparison Diagnosis Model

The (n, k)-star graph, denoted by, S-n,S-k is an enhanced version of n-dimensional star graphs, that has better scalability than S-n, and possesses several good properties, compared with hypercubes. Diagnosis has been one of the most important issues for maintaining multiprocessor-system reliability. Conditional diagnosability, which is more general than classical diagnosability, measures the multiprocessor-system diagnosability under the assumption that all neighbors of any processor in the system cannot fail simultaneously. In this paper, we investigate the conditional diagnosability of S-n,S-k for (n >= 3 and k = 1) and (n >= 4 and 2 <= k <= n) under the comparison diagnosis model.