Multi-granularity Software Rejuvenation Policy Based on Continuous Time Markov Chain

Abstract

In this paper, a multi-granularity software rejuvenation policy isstudied. Four granularities of rejuvenation are proposed tomitigate the impact of four levels of software aging respectively.Continuous Time Markov Chain (CTMC) model is used to obtain theavailability and cost from its closed form solution. Moreover, therelations between availability, cost and parameters are studied and the optimal selection of parameters is obtained accordingly. Numerical examples show that the optimal inspection rates for maximizing availability and minimizing cost become larger if inspection delay rate becomes lower, and also the optimal inspection rates have the property of phase transitions when inspection delay rate changes.