影响Web服务组合性能的因素分为“内因”和“外因”,内因具体表现为:BPEL流程的结构、BPEL流程中变量取值的概率分布；外因具体表现为:Web服务器的软硬件处理能力、Web服务器的负载(包括“正对其进行性能分析的Web服务组合”对Web服务器形成的负载和“其它Web服务组合”对Web服务器形成的负载)以及Web服务器的调度策略.目前广泛采用的广义Petri网、排队Petri网、Markov过程和随机进程代数等模型不能同时综合建模上述各种“内因”和“外因”对Web服务组合性能的影响,导致不能全面分析Web服务组合在互联网环境下的性能.文中建立了一组把影响Web服务组合性能的各种“内因”和“外因”映射到具有G/G/1-FCFS、M/G/1-PS和M/G/∞排队节点类型的排队网络的映射规则,给出了一组建立在排队网络基础上的Web服务组合性能分析指标体系及其计算公式,并以这些性能分析指标体系为基础,分析了Web服务组合的性能及其变化规律,以便在Web服务组合部署前,分析预测Web服务组合在互联网环境下的性能.%Elements affecting the performance of Web services composition consist of internal and external factors. The internal factors include the structure of the BPEL process, the probability distribution of the variables of BPEL process. The external factors include the processing capabilities of hardware and software of Web servers, the load of Web servers, and the scheduling policy of Web servers. The generalized Petri nets, queuing Petri nets, Markov processes and stochastic process algebra model can not comprehensively establish the performance model of Web services composition employing these various internal and external factors at the same time, resulted in lack of obtaining the actual performance of Web services composition in the Internet environment. In this paper, a set of mapping rules is established to transform the internal and external factors into queuing network with G/G/1-FCFS, M/G/1-PS, and M/G/oo node types, and a set of performance indexes and its formula is given to evaluate the Web services composition. Based on these performance indexes, we can comprehensively analyze the performance of Web services composition and predicate its performance before its deployment.
展开▼
