A Petri-Net Based Reliability Prediction Method for SOA Software

In order to satisfy the requirement of the use of large-scale, complex distributed software system, Service Oriented Architecture has aroused wide concern. How to build a reliable SOA software system has become one of the important issues in this research area. As for SOA software, reliability prediction can be used for providing reliability information at the early stage of development to help developers in improving the quality of software. In current researches of SOA reliability prediction, on the one hand, researchers mainly focus on only one aspect of software system like service processes or service states, neglecting a comprehensive view of both static structure and dynamic behavior. On the other hand, the participation of users is not fully considered and modeled. In our previous research, we proposed a SOA software architecture extended modeling method considering reliability information. In this paper, we continued our research and a SOA software reliability prediction method is proposed. The proposed approach is based on our previous modeling method, and a model transformation approach is used to build the Petri-net based reliability analytic model. Then the reliability prediction result is obtained by Petri-net simulation. The proposed software prediction method can provide reliability information at the early stage of software development, which is helpful for ensuring the reliability requirement and the quality of software at the end of the development.