Rough Set-Based Debugging for Web Services System

Web services technology provides a flexible and cost-effective paradigm to construct highly dynamic systems through service discovery, composition, and ultra-late binding. However, its new features bring great pressure to maintain Web service-based system. Based on the massive testing results, how to locate the fault points in system is a challenging task. In the paper, a two level diagnosis framework for Web services system is proposed. In service unit level, the WSDL interface information is used to construct decision table. In service composition level, the decision information system is built by comprehensively using process specifications and interface information. Then, rule mining algorithm in rough set reasoning is adopted to reveal the input cases associated with service or system failures. How to utilize such rules to locate faults in Web services system is also discussed. In addition, two cases are introduced to validate the feasibility and effectiveness of our approach.