An execution context aware approach for Web service substitution

Web service substitution refers to the problem of identifying an optimal service that can replace another service for the same abstract service in a composition. However, the quality of Web services is closely related to their execution context. For example, a Web service will have different performance if it is invoked in different network conditions, transferring and processing different data amount, or in different service host conditions. Most existing service substitution approaches ignore this issue so as to make the substituted composite service inefficient or even unexecutable. To solve this problem, an execution context aware approach for Web service substitution is proposed in this paper. Firstly the execution context data of composite services is recorded into log repositories. Then, the Web service execution context feature rules which qualitatively express the different performance of Web services in different execution context conditions are mined. At last, the discovered knowledge is applied to choose the optimal candidate substitutive service according to current execution context conditions. Extensive experimental results show that our approach can improve the robustness and efficiency of substituted composite services.