Permission-based fault-tolerant distributed mutual exclusion algorithm

Fault tolerance is an important issue in a distributed computing environmen t. Without a suitable fault-tolerant algorithm, a node failure would lead t o total system failure. In this work, a permission based fault-tolerant mut ual exclusion algorithm for distributed systems is proposed. The algorithm detects node failures and automatically isolates the failed nodes. Hence re st of the system continues to function normally. The proposed algorithm use s both dynamic and static information structures. The dynamic information e volves as and when node failures occur in the system. The novelty of using the dynamic information structure helps in overcoming the drawbacks of the fault-tolerant distributed mutual exclusion algorithms presented by Bouabda llah and Wong et, al. In the present algorithm the message complexity remai ns the same in the absence of failure, which is an advantage in contrast to the algorithm by Bouabdallah. The algorithm also takes care of all possibl e failures which affect the system including the cases that are not being h andled properly by Wong et al. Over and above these advantages, the present algorithm enshrines a restart facility to re-configure a failed node back into the system as and when it comes up after rectification. The correctnes s of algorithm is verified by showing that the algorithm satisfies both saf ety and liveness properties. Simulation studies, done on sample systems, es tablish the practicability of the proposed algorithm.