面向内容发布订阅系统的混合事件匹配算法

Hybrid Event Matching Algorithm for Content-Based Publish/Subscribe System

当前的事件匹配算法不能在高效匹配的同时满足频繁订阅变更的要求.在结合已有谓词索引算法和覆盖网络算法的基础上,融合谓词索引结构的易变更和覆盖网络的高效匹配特点,提出一种混合的事件匹配算法.算法将部分订阅覆盖关系从覆盖网络中剥离,以同谓偏序的形式引入到谓词索引结构中去,达到高效匹配的同时保留了谓词索引的易变更结构.实验表明,与同类算法相比该算法能够在频繁订阅情况下提供高效的匹配,从而满足相关应用的需求.

Current typical content-based publish / subscribe systems are not efficient in subscription processing or event matching. This paper presents hybrid event matching algorithm（ HEMA）,a novel publish / subscribe systems which joins predicate indexing and testing network approaches. We put partially ordered subscription with same predicates,which are separated from testing network structures,into predicate indexing mechanism to sustain efficient matching,whilst changing large number of subscriptions. Finally,experiments and performance analysis showthat HEMA significantly improve throughput of event propagation and reduce response time to subscription updates meanwhile.