基于PEPA的云计算资源分配算法性能评价  

PEPA model approach for performance evaluation of dynamic resource provision in cloud computing

作　　者：王倩[1] 石振国[2] 孙万捷[2] 刘赛男[1] 王国涛[1] 

机构地区：[1]南通大学电子信息学院,江苏南通226019 [2]南通大学计算机科学与技术学院,江苏南通226019

出　　处：《计算机应用研究》2015年第4期1179-1183,1187,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(60975033);江苏省高校自然科学基金资助项目(07KJB520096)

摘　　要：资源分配是云计算的核心之一,对云计算资源分配算法的性能进行评价可为云计算平台设计提供指导。讨论了两种云计算资源分配算法,提出了一种基于PEPA的资源分配算法的性能评价模型,该模型通过建立云计算系统中各组件之间的交互关系进行形式化分析和推理,获得了云计算系统性能的评价指标。实验通过分析资源分配过程中不同参数变化对系统性能的影响,结果表明,PEPA模型方法可以直接评估资源分配算法性能的优劣,并能够确定算法性能提升的关键因素,从而减少云平台设计过程的周期。One of the significant issues in cloud computing is the virtualized resources provision,since it will affect the performance of the entire system. Hence,performance evaluation of dynamic resources provision is necessary to provide some design guild lines. This paper adopted a method for analyzing the performance of resource provisioning process using PEPA approach. It analyzed two different PEPA models and then obtained the performance evaluation index by solving the underlying continuous time Markov chain（ CTMC）. Experimental results indicate that efficiency of resources provision algorithm determines the performance of the cloud computing system primarily,and the performance varies diversely under distinctive circumstances,which can be referenced for finding optimal strategy in cloud computing platform designing and deploying.

关 键 词：PEPA(性能评价进程代数) 云计算 资源分配算法 性能评价 

分 类 号：TP393.07[自动化与计算机技术—计算机应用技术]