Abstract：In order to handle software dependability crisis, dependability growing method of software based on self-configuration was proposed, which enables the software to have selfconfiguration capabilities and enhance the software dependability. Using self-configuration mechanism of autonomic computing, three self-configuration policies were defined and a software dependability growth model based on self-configuration (SDGMS) was established. SDGMS was analyzed by means of Markov regenerative stochastic Petri net (MRSPN) and related parameters in steady state was obtained. Software users were simulated by JMeter through launching HTTP requestes. The related parameters were compared between the software improved by the method and original one. The results show that the method can increase software dependability and decrease software maintenance cost by selecting the appropriate software self-configuration period. The method is helpful for dependability growing software and can improve the computer system running the software.
2010, Vol. 31 
