A study on software rejuvenation model of application server cluster in two-dimension state space using Markov process

Software rejuvenation is a main method for counteracting software aging, when it is performed in cluster system, system reliability can be significantly increased. We firstly presented a software rejuvenation policy of application sever cluster. Then we set up a software rejuvenation model in two-dimension state space for a two-node application server cluster and give the quantitative analysis using continuous time Markov process. Finally, the numerical experiment results show that the selected optimal software rejuvenation interval can greatly reduce expected systematic downtime and improve software reliability of application server cluster. © 2008 Asian Network for Scientific Information.