Results on Out-of-Order Event Processing

Abstract

Complex Event Processing (CEP) has the task of processing streams of events with the goal of detecting event patterns of interest. Today’s CEP systems typically assume the total order of streaming events. In practice, real-time processing often faces delays caused by network latencies, sensor and machine failures etc. By handling out-of-order events a CEP processor needs to keep certain events longer than they are normally needed (in order to handle late events). Therefore, an effective removal of overdue events is needed. This work provides a framework for processing events, including also out-of-order events. The framework also implements a general low-level garbage collector. Our approach is based on deductive rules where detection of complex events amounts to an inference procedure. Therefore the framework features reasoning capabilities. We provide an open source implementation, and present experimental results of the proposed framework.