OPTIMAL SOFTWARE REJUVENATION FOR TOLERATING SOFT FAILURES

In recent studies, the phenomenon of software ''aging'' has come to li ght which causes performance of a software to degrade with time. Softw are rejuvenation is a fault tolerance technique which counteracts agin g. In this paper, we address the problem of determining the optimal ti me to rejuvenate a server type software which experiences ''soft failu res'' (witnessed in telecommunication systems) because of aging. The s ervice rate of the software gradually decreases with time and settles to a very low value. Since the performability in this state is unaccep table, it is necessary to ''renew'' the software to its peak performan ce level. We develop Markov decision models for such a system for two different queuing policies. For each policy, we define the look-ahead- n cost functions and prove results on the convergence of these functio ns to the optimal minimal cost function. We also prove simple rules to determine optimal times to rejuvenate for a realistic cost criterion. Finally, the results are illustrated numerically and the effectivenes s of the MDP model is compared with that of the simple rules.