基于事件列表递归比较的复杂事件匹配方法

夏秀峰; 丁建丽; 邱涛; 谢沛良

doi:10.3969/j.issn.2095-1248.2023.02.005

沈阳航空航天大学学报 ›› 2023, Vol. 40 ›› Issue (2) : 31 -43. DOI: 10.3969/j.issn.2095-1248.2023.02.005

信息科学与工程

基于事件列表递归比较的复杂事件匹配方法

作者信息 +

Complex event matching based on recursive comparison of event list

Author information +

文章历史 +

PDF (1627K)

摘要

针对现有复杂事件匹配处理方法存在匹配代价高的问题，提出了一种在有序事件列表上选择最佳匹配顺序进行递归遍历的复杂事件匹配方法OptiSeq。将事件实例按照查询模式中不同事件类型缓存到有序事件列表中，并通过事件列表中事件实例的数量选择最优的查询匹配起点及查询匹配顺序，之后在有序列表上对不同约束分别进行递归校验，最终输出完全满足查询模式的所有复杂事件结果。该方法克服了使用自动机模型固定状态转换的弊端，也避免了使用树型模型批处理操作漏解的问题，并且合理优化了匹配顺序，进一步提高查询匹配效率。在模拟数据集和真实数据集上进行了实验测试与分析，与当前流行的SASE、Siddhi两种复杂事件处理方法进行比较。实验表明，所提出的方法能够在保证匹配正确性的同时，有效地减少匹配过程中的冗余计算，提高整体匹配效率。

Abstract

Aiming at the problem of high matching cost in the existing complex event matching processing methods， a complex event matching method OptiSeq was proposed， which selected the best matching order on the ordered event list for recursive traversal. The event instances were cached in the ordered event list according to the different event types in the query mode， and the optimal query matching starting point and query matching order were selected by judging the number of event instan ces in the event list. Then， the OptiSeq method performed recursive validation and ultimately output all complex event results that fully met the query pattern.This method overcame the drawbacks of using automata model to fix state transitions， and also avoided the problem of missing solutions in batch operations using tree models， and optimized the matching order reasonably， further improved query matching efficiency. Experiments were tested and analyzed on simulated datasets and real datasets， and compared with the current popular SASE and Siddhi complex event processing methods. The experiments show that the proposed method can effectively reduce the redundant calculations in the matching process and improve the overall matching efficiency while ensuring the correctness of matching.