On the consistency of event processing: A semantic approach

Event processing is one of the cornerstone technologies in bridging physical world and cyber system together. Although event-based processing system has been widely used in various applications, however, consistency of event processing is still an open issue need further exploration. The inconsistency problem produces inaccurate detection result and corrupts the system correctness. In this paper, we propose a semantic approach for event modeling and detection model transformation with a semantic calculus system. Particularly, we first propose a complex event semantic model, OntoEvent, and define several key operators and properties to describe the logic, temporal and attribute relations in complex events. Second, we propose the concept of event constraint and elaborate the occurrence, temporal, and attribute functions to formalize the semantic implications in OntoEvent model. On that basis, we present the extraction rules and establish a calculus mechanism for constraints based on axioms. With these works, an automata-based detection model, named OntoCEP (Ontology-based Complex Event Detection), and a pipelined procedure for the assembly from constraints to OntoCEP model is proposed. The procedure is composed of several sequential phases and the consistency in each assembly phase is proved. Therefore, we establish a semantic-consistent mapping mechanism from event to detection model in the form of constraints. Experiments and evaluations prove that our approach ensure the consistency with event and detection models. Besides, our detection model consumes less computational resources and outperforms other selected benchmarked models in terms of computational efficiency and processing capability.