一种基于事件驱动的SOA故障疑似集选择算法

摘要

The flexibility and dynamic property of service-oriented architecture (SOA) make services behaviours at runtime of monitoring and managing critical to the performance assurance. This paper proposes an event-driven based fault suspected-set selection (FSS) algorithm of SOA according to Bayesian fault diagnosis network. The algorithm integrates the sensitivity analysis technology in Bayesian Network theory and the k-median model, then adds the fault identifying set, selects the corresponding fault suspected-set according to the specific fault event.Simulation results show that the algorithm has higher fault suspected-set search completeness rate. The added fault identifying set is also conductive to the future predictive analysis.%面向服务体系架构(SOA)的灵活性和动态性,使得监测和管理运行时服务行为成为性能保证的关键所在.依据贝叶斯故障诊断网络提出了一种基于事件驱动的SOA故障疑似集选择FSS(Fault Suspected-set Selection)算法,该算法综合贝叶斯敏感性分析技术以及k-median模型,并加入故障标识集,根据具体的故障事件选择对应的故障疑似集合.仿真实验表明,该算法具有较高的故障疑似集查找完整率.增加的故障标识集也有利于以后的预测性分析.