计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (8): 68-75.DOI: 10.3778/j.issn.1002-8331.1510-0167

• 大数据与云计算 • 上一篇    下一篇

分布式复杂事件流处理平台的研究

孔祥翾1,廖湖声2,王亦雄1,高红雨1,苏  航1   

  1. 1.北京工业大学 计算机学院,北京 100124
    2.北京工业大学 软件学院,北京 100124
  • 出版日期:2017-04-15 发布日期:2017-04-28

Distributed complex event stream processing platform research

KONG Xiangxuan1, LIAO Husheng2, WANG Yixiong1, GAO Hongyu1, SU Hang1   

  1. 1.College of Computer Science, Beijing University of Technology, Beijing 100124, China
    2.Software College, Beijing University of Technology, Beijing 100124, China
  • Online:2017-04-15 Published:2017-04-28

摘要: 复杂事件处理是大数据时代的关键技术之一。针对现有的复杂事件处理平台存在的各种问题,提出并设计了一个以复杂事件流处理语言CEStream为基础的分布式复杂事件流处理平台,实现了基于正规树模式的事件检测功能,支持多数据源的复杂事件处理。针对多数据源组合事件模式的检测需求,该平台可以将复杂事件分解为多个独立的事件检测任务,部署在集群中不同节点和远端的事件检测代理上。实验结果表明,系统的吞吐量和响应时间达到了预期结果,可以满足目前主流的复杂事件处理的应用需求。

关键词: 复杂事件处理, 流式计算, 分布式系统, 正规树模式

Abstract: Complex event processing is one of the core technologies in?the big-data age. Since existing several?problems in complex event processing platform, a distributed complex event stream processing platform based on CEStream(a complex event stream processing language) is proposed, providing event detection based on regular twig pattern and complex event processing of multi-sources. According to composite event pattern of multiple sources, the platform can decompose complex event into several?independent?event detection?tasks and deploy them on different nodes of cluster and remote event detection agents. As the experiments show, the throughput and response time of the system comply with the requirements of mainstream of complex event stream processing.

Key words: complex event processing, stream computing, distributed system, regular twig pattern