Modeling and Analysis of Load and Time Dependent Software Rejuvenation Policies

Due to repeated and potentially faulty usage of continuously running client-server type software systems by many clients, such software \ages" with time and eventually fails. Huang et. al. proposed a technique called \software rejuvenation" [3] in which the software is periodically stopped and then restarted in a \robust" state after proper maintenance. This \renewal" of software prevents, or at least postpones, the crash failure. In this paper, we present a quantitative analysis of two software rejuvenation policies. The rst one considers only the ageing behaviour of the system by time, while the second one considers the actual load of the system as well. The behaviour of the system is represented through a Markov Regenerative Stochastic Petri Net (MRSPN) model. Numerical analysis of the system performance regarding the probability of successful service of clients is provided.