A new algorithm for k mutual exclusions in distributed systems

Abstract

The speedup achieved by parallel execution in a distributed system is hindered when only one process is allowed to enter the critical section to access a shared resource. In certain cases, distributed systems can provide duplicated resources that allow multiple processes to execute the critical section simultaneously while maintaining consistency within the system. This article presents a new, more efficient algorithm that allows multiple simultaneous entries to a critical section in a distributed system. Correctness proofs are given and theoretical best- and worst-case performances analyzed. Extensive simulation results are provided to evaluate the algorithm's average case performance. It is shown that the proposed algorithm achieves logarithmic communication efficiency (in terms of the number of messages exchanged per critical section invocation) compared with linear communication efficiency of all similar existing algorithms.