Reliability Analysis of Component-Based Software Based on Relationships of Components

Recent development in daily business life indicates demand for higher reliability of SOA software. After analyzing the traditional software reliability models (Markov chain based), it was found that many software with complicated relationships of components did not satisfy the Markov property. This paper analyzed various complex components´ relationships (parallel, loops, backup, fault- tolerance, request and response, etc.) how to influence the reliability of the whole system, and solved these complicated relationships to satisfy the Markov property. This work extended the scope of the application of the Markov Model and improved the accuracy of reliability estimation.