A Monte Carlo simulation based dependability analysis of a non-Markovian grid computing environment with software rejuvenation

A grid system with star topology is considered consisting of a Resource Management System (RMS) and n distributed Root Nodes (RNs). It is also assumed that the RMS suffers from software aging phenomena and software rejuvenation is adopted to counteract them. Our aim is to examine the grid reliability, availability and performance, in the case that the lifetimes of all the nodes are exponentially distributed while their repair times follow a general distribution. Although the evolution in time of each component of the system is described by a continuous time semi-Markov process, this is not the case for the whole system. Usually in such a system, an analytical computation of the dependability measures is very difficult or even infeasible. Consequently Monte Carlo methods are a common approach for studying these systems. Thus, the system's behavior in time is modeled and simulated. The resulting output is used for estimating the principal dependability and performability measures.