Entity-Based State Management For Complex Event Processing Applications

Complex Event Processing (CEP) using Event-Condition-Action (ECA) rules has proved particularly suitable for detecting noteworthy business situations of a defined length and structure. By contrast, challenges arise when the state of a complex, durable entity - e.g., a counter, a server, or a task queue - shall be derived from continuous streams of low-level updates. In this paper, we present a novel approach to state management for CEP applications. We propose business entity providers, which encapsulate arbitrary state-calculation logic and manage state in the form of typed, application-wide data structures. Using a plug-in-based component model, business entity providers can be integrated into an application based on the specific requirements of a business scenario. We present an ECA rule model that allows accessing business entities well-integrated with event-pattern detection and demonstrate our approach in a real-world scenario from the workload automation domain.