Process mining has emerged as a way to discover or check the conformance of processes based on event logs. This enables organizations to learn from processes as they really take place. Since web services are distributed over autonomous parties, it is vital to monitor the correct execution of processes. Fortunately, the "web services stack" assists in collecting structured event logs. This information can be used to extract new information about service processes (e.g., bottlenecks, unused paths, etc.) and to check the conformance (e.g., deviations from some predefined process). In this paper, we illustrate the potential of process mining in the context of web services. In particular, we show what a process mining tool like ProM can contribute in IBM's WebSphere environment.
Wil M. P. van der Aalst, H. M. W. (Eric) Verbeek