Managing Measurement and Occurrence Uncertainty in Complex Event Processing Systems

Complex event processing (CEP) is a powerful technology for analyzing streams of real-time events, coming from different sources, and for extracting conclusions from them. In many situations, these events are not free from uncertainty, due to either unreliable data sources and networks, measurement uncertainty, or inability to determine whether an event has actually happened or not. This paper presents a proposal for incorporating and managing different kinds of uncertainty that may happen in both events and rules of the CEP systems. We provide a library that enables the representation and propagation of uncertain values, which can be efficiently integrated with the existing CEP languages and engines to deal with uncertainty, and we show how the treatment of uncertainty can be smoothly added to two of them: Esper and Apache Flink. Five applications coming from various domains serve to evaluate the proposal and to analyze its performance and accuracy. The results show that the overhead introduced by the treatment of uncertainty is not high and good precision and recall are achieved.