The Relationship between Connectivity and Diagnosability on Regular Networks

The node connectivity is usually used to measure the fault tolerance of a multiprocessor system. On the other hand, the diagnosability is an important ability to provide a system level self-diagnosis for maintaining the reliability of a multiprocessor system. It is interesting to investigate the relationship between the connectivity and the diagnosability. Two diagnosis models: PMC model and comparison model are studied in this work. Herein, we consider that, under certain condition, a t-connected network is a t-diagnosable and is also a strongly t-diagnosable network. From our results, we can conclude that the diagnosability is enhanced as the connectivity improves for a multiprocessor system.