Towards Automating the Detection of Event Sources

During business process execution, various systems and services produce a variety of data, messages, and events that are valuable for gaining insights about business processes, e.g., to ensure a business process is executed as expected. However, these data, messages, and events usually originate from different kinds of sources, each specified by different kinds of descriptions. This variety makes it difficult to automate the detection of relevant event sources for business process monitoring. In this paper, we present a course of actions to automatically associate different event sources to event object types required for business process monitoring. In particular, in a three-step approach we determine the similarity of event sources to event object types, rank those results, and derive a mapping between their attributes. Thus, relevant event sources and their bindings to specified event object types of business processes can be automatically identified. The approach is implemented and evaluated using schema matching techniques for a specific use case that is aligned with real-world energy processes, data, messages, and events.