Generating Quantitative Test Cases for Probabilistic Timed Web Service Composition

Abstract

The environment of enterprise applications is characterized by frequently changing market demands, time-to-market pressure and fierce competition. To seamlessly integrate complex computing activities, Web Service Composition (WSC) has been regarded as an emerging E-Commerce solution to support interoperable machine-to-machine business interactions over network. To guarantee the composite Web service can be successfully produced, testing is a preferred validation technique to efficiently verify the correctness of functional and nonfunctional requirements of WSC behaviors. BPEL4WS is a high level and semi-formal abstract description language for WSC orchestration. Manually generating test cases from BPEL4WS is tedious, time-consuming, and error prone. Thus, the automated test case generation plays a critical role in all the phases of Web service life cycle. Considering the uncertain environment, an extended WSC model, namely probabilistic timed interface automata for Web service (PTIA4WS), is pro-posed to transform and extend BPEL4WS with regard to the stochastic and time-related behaviors. Based on PTIA4WS model, we propose an approach for generating quantitative test cases from counterexamples of violated PTCTL formulae using coverage criterions. After that, timed test case with fastest execution response time and probabilistic test case with maximal execution success rate are discussed. The series of experiments show that our method gains better performance than traditional methods.