基于排队论的云计算资源池模型研究  

Computing Clouds Resources Pool Model Research Based on Queue Theory

作　　者：刘赛[1] 李绪蓉[1] 万麟瑞[1] 

机构地区：[1]南京航空航天大学,江苏南京210016

出　　处：《计算机技术与发展》2012年第12期87-89,共3页Computer Technology and Development

基　　金：国家"863"高技术发展计划项目(86351181004103)

摘　　要：云计算环境中资源池管理是云计算的关键技术之一。文中通过在用户作业使用云计算资源服务时间满足负指数分布的情况下,利用马尔可夫链和排队论的相关性质方法,构造了云计算资源池多服务窗等待排队模型M/M/n。阐述了在给出一个排队队列时和多个排队队列时的系统工作性能比较,以及资源池中服务进程与作业等待时间对比关系。论证了云计算资源池模型的合理性,得出在作业忍受一定排队等待时间的条件下,云计算资源池集成适当数量的资源,就可以为用户提供较高性价比的云计算服务。最后探讨了进一步的研究方向。Cloud computing resource pool management is one of the key technologies in cloud computing environment. In this paper,the cloud computing resources in the user＇s job to meet the circumstances of the negative exponential distribution, using the relative nature method of the Markov chain and qneueing theory,constructed the wait queueing model M/M/n of cloud computing resource pool service window： Described the performance contrast of system work when given a queue and multiple queue and comparsion relation between service process and operation waiting time in the resource pool. The rationality of a cloud computing resource pool model is proved, and come to tolerate the job under certain conditions queued time, the cloud computing resource pool to integrate the appropriate number of resources, can provide users with a higher cost of cloud computing services. Finally, discuss the future research directions.

关 键 词：云计算 资源池 M M N 排队论 马尔可夫链 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]