Fault Recovery in Distributed Processing Loop Networks

A graph model is introduced to formalize the fault recovery process in distributed loop networks. This model is applicable to centralized as well as distributed recovery. Key fault tolerance and recovery parameters including redundancy, fault model, recovery time, and recovery strategy are characterized. Centralized recovery strategies for a given fault-tolerant loop network are presented and analyzed. A distributed recovery strategy, which depends on the cooperation of a set of processors, is given, and its application to a new class of fault-tolerant loop networks is evaluated.