一种基于随机Petri网的Web服务组合性能瓶颈定位策略

发布时间：2018-04-13 21:05

本文选题：随机Petri网 + Web服务　；参考：《计算机学报》2013年10期

【摘要】：Web服务组合的性能对推动以服务为核心的价值链的良性融合起着重要作用.针对如何有效发现Web服务组合中性能瓶颈的问题,提出一种基于随机Petri网的Web服务组合性能分析模型,根据库所和变迁之间的直接关系总结出模型的4种基本结构,通过结构转换得到性能瓶颈定位分析结构集{顺序结构,分支结构,汇聚结构},在此基础上提出一种Web服务组合性能瓶颈定位策略.该策略首先根据库所中token消耗的速度将库所分成主动受阻库所、被动受阻库所和流通库所3类,然后分别基于3类库所在3种分析结构中的不同分布情况,通过顺向结构分析和逆向结构分析相结合的方法,讨论瓶颈可能存在的位置,并给出如何确定3种分析结构的结构发现定理.实验证明策略的预测结果是有效的.
[Abstract]:The performance of Web services composition plays an important role in promoting the good integration of service-based value chain.To solve the problem of finding performance bottleneck in Web service composition effectively, a performance analysis model of Web service composition based on stochastic Petri net is proposed. According to the direct relationship between library and transition, four basic structures of the model are summarized.The structure set {sequence structure, branch structure, convergence structure} is obtained by structure transformation. Based on this, a performance bottleneck location strategy for Web services composition is proposed.According to the speed of token consumption in the library, the strategy divides the library into three categories: active blocked library, passive blocked library and circulation library. Then, the strategy is based on the different distribution of the three kinds of structures in which the library is located.By combining forward structure analysis with reverse structure analysis, the possible location of bottleneck is discussed, and how to determine the structure discovery theorems of three kinds of analytical structures is given.Experiments show that the prediction results of the strategy are effective.
【作者单位】：武汉大学计算机学院;武汉大学软件工程国家重点实验室;湖北工业大学计算机学院;
【基金】：国家“八六三”高技术研究发展计划“动态Web服务组合关键技术及其性能分析研究”(2007AA01Z138) 国家自然科学基金重点项目“可信软件构造理论与方法研究”(91118003) 国家自然科学基金面上项目“基于编译的嵌入式系统优化研究”(61170022)资助~~
【分类号】：TP393.09

【参考文献】