基于云计算集群扩展中的调度策略研究  

Scheduling strategy research based on cloud computing for cluster expansion

作　　者：葛新[1] 陈华平[1] 杜冰[1] 李书鹏[1] 

机构地区：[1]中国科学技术大学管理学院,合肥230026

出　　处：《计算机应用研究》2011年第3期995-997,1006,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(70821001);国家博士点基金资助项目(200803580024)

摘　　要：为了达到更为经济的扩展集群的目的,从使用云计算扩展集群的角度出发,提出了一种集群扩展的基本构架;描述了本地优先策略和云环境优先策略,并根据云计算的特点,提出了一种基于计算强度的新型调度策略。此外,设计了调度器系统,并在不同环境下对每种策略进行了分析比较,得出基于计算强度的调度策略的低成本和较多任务超时的特点。最后,对如何使用云计算扩展本地集群更为经济进行了分析。For economic expansion of cluster,this paper pointed out a basic framework based on the capability of cloud computing to increase the capacity of clusters.It discussed two scheduling strategies,i.e.,local priority strategy and cloud environmental priority strategy and proposed a new scheduling strategy based on calculation.Also,designed scheduler system indicated that the proposed scheduling strategy would require less cost but possess more overdue tasks other than the two strategies under several conditions.Finally,concluded how to use cloud computing more economically to extend the local cluster.

关 键 词：云计算 集群扩展 虚拟化 调度 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]