摘要
Web服务组合系统的规模尺度和复杂程度的不断提高以及调用请求的周期化和一般化,使得Web服务组合系统的性能建模和分析更加困难。针对该问题,建立了一组将服务组合的基本交互关系和Web服务器处理能力、负载等因素映射到开排队网络的映射规则。提出了一种基于逼近理论与排队模型相结合的分析方法,克服了以往研究中调用请求到达过程和服务过程受限的情况,得到了系统关键性能指标的流逼近与扩散逼近定理。最后,分析了系统的性能及其变化规律。该方法为研究更加一般的大型Web服务组合系统提供了一个新的途径。
With the improvement of the Web service composition systems’size and complexity,the periodism and generalization of invoking requirements,the modeling and analyzing of systems become more difficult.To tackle this problem,this paper established a set of mapping rules to transform the interaction relations and factors including the processing capabilities and the load of Web servers into open queueing network.It developed a new analytic technique based on approximation theories and queueing networks,overcame the restriction of arrival processes and service processes for invoking requirements in past studies and got the fluid and diffusion limit theorems of some key performance indexes.At last,this paper analyzed the performance and its change rules of systems.The proposed results and methodology provide new highlight in the study of more general large-scale Web service composition systems.
作者
骞志勇
樊瑞娜
Qian Zhiyong;Fan Ruina(School of Economics&Management Sciences,Yanshan University,Qinhuangdao Hebei 066044,China)
出处
《计算机应用研究》
CSCD
北大核心
2018年第3期795-802,共8页
Application Research of Computers
关键词
WEB服务组合
排队网络
流极限
扩散极限
性能分析
Web service composition
queueing networks
fluid limit
diffusion limit
performance analysis