Fault diagnosis analysis of matching composition networks

The growing size of the multiprocessor system increases its vulnerability to component failures. It is crucial to locate and to replace the fault processors to maintain a system´s high reliability. The fault diagnosis is the process of identifying faulty processors in a system through testing. Based on the fault tolerance of the matching composition networks, this paper proposes a fast pessimistic diagnosis algorithm on the matching composition networks.