An efficient distributed deadlock avoidance algorithm for the AND model

Abstract

A new rank-based distributed deadlock avoidance algorithm for the AND resource request model is presented. Deadlocks are avoided by dynamically maintaining an invariant Con(WFG): For each pair of processes pi and pj, pi is allowed wait for process pj iff the rank of pj is greater than that of pi for the WFG (Wait-For Graph). Our algorithm neither restricts the order of resource requests nor needs a priori information about resource requests nor causes unnecessary abortion of processes. Multidimensional ranks, which are partially ordered and dynamically modified, are used to drastically reduce the cost of maintaining Con(WFG). Our simulation results show that the performance of our algorithm is better than that of existing algorithms.