BSFDiag: Behavior Similarity-based Fault Diagnosis for Web Services ⋆

With the development of web service technology, web service becomes a promising solution for satisfying the requirements of real-time transactions on Internet. However, there still exist many problems to cause web services to throw the exceptions or halt. Hence, the real-time fault diagnosis is a critical issue for web services. In this paper, we firstly present a diagnostic architecture with hierarchical scalability to enhance the diagnostic capability and guarantee the privacy of web services. And then we propose a behavior similarity-based diagnosis method to improve the diagnostic accuracy. Our method analyzes the similarity between behavior dependence relations and exception executions to compute the suspiciousness score of each behavior, and localizes the behavior with the maximum score as the fault. Finally, we provide a case study to illustrate the diagnostic process and conduct the simulation experiments to evaluate our method. The experimental results show that our method is effective and asymptotically optimal with increasing execution information.