An intelligent self-adaption complex event processing framework with dynamic context detection and automatic event pattern modification abilities

Abstract

Many application domains involve dynamic context situations and the problem of how to adapt to them. This requires a Complex Event Processing (CEP) engine specifically designed to have a self-adapting ability to timely adjust its behaviors according to changing circumstances. To meet this need, this paper proposes AC-CEPs, an adaptive control system framework that can monitor context, make business decisions and adjust itself automatically. This framework not only provides a SCEP rule language able to declare event’s semantics, but also integrates complex event detection, semantic annotation and semantic inference to detect dynamic context. And a two-library intelligent decision support system based on Drools is used to calculate and decide how to adapt to the targets. XML configuration file is also used to describe the event pattern architecture to support flexible and fast event pattern modification. In addition, an event parameter pre-filter net is designed to optimize adapting efficiency. Experiments showed that AC-CEPs’ modules could operate in a coordinated way under different workloads, demonstrating the feasibility of this framework.