Design and Implementation of A Distributed Complex Event Processing System

摘要

Making use of the massive events from event sources such as sensors and bank transactions and extract valuable information is of significant importance. Complex Event Processing (CEP), a method of detecting complex events from simple events stream, provides a solution of processing data in real time fast and efficiently. However, a single node CEP system can't satisfy requirements of processing massive event streams from multitudinous event sources. Therefore, this article designs a distributed CEP system, which combine Siddhi, a CEP engine, and Storm, a distributed real time computation architecture. This system can construct topology automatically based on the event streams and execution plans provided by users and process the event streams parallel. Compared with single node complex event system, the distributed system can achieve better performance.