Efficient analysis of event processing applications.

ABSTRACTComplex event processing (CEP) middleware systems are increasingly adopted to implement distributed applications: they not only dispatch events across components, but also embed part of the application logic into declarative rules that detect situations of interest from the occurrence of specific pattern of events. While this approach simplifies the development of large scale event processing applications, writing the rules that correctly capture the application domain arguably remains a difficult and error prone task, which fundamentally lacks consolidated tool support. Moving from these premises, this paper introduces CAVE, an efficient approach and tool to support developers in analyzing the behavior of an event processing application. CAVE verifies properties based on the adopted CEP ruleset and on the environmental conditions, and outputs sequences of events that prove the satisfiability or unsatisfiability of each property. The key idea that contributes to the efficiency of CAVE is the translation of the property checking task into a set of constraint solving problems. The paper presents the CAVE approach in detail, describes its prototype implementation and evaluates its performance in a wide range of scenarios.