基于代价感知的云任务处理策略  

Cost-Aware Workload Process Strategy in Cloud Environment

作　　者：况亚萍[1] 王雷[1] 刘小龙[1] 黄承真[1] 

机构地区：[1]中国科学技术大学自动化系网络传播系统与控制安徽省重点实验室,合肥230039

出　　处：《计算机系统应用》2013年第11期148-153,共6页Computer Systems & Applications

基　　金：中央高校基本科研业务费专项基金(WK2100100012)

摘　　要：绿色云计算旨在保证服务质量的前提下,最大限度地降低云服务成本.针对这一问题,提出一种兼顾云数据中心高性能和低成本的策略,通过虚拟机动态迁移等技术进行云任务聚合,关闭空闲的节点以最小化能量消耗,同时考虑虚拟机迁移性能,以此保证任务的服务质量.实验结果表明,代价感知的策略可以减少打开节点的个数,兼顾迁移能耗和迁移延迟,在节能的前提下使服务质量得到保证.Green cloud computing aims at reducing the cost of cloud services in the premise of guaranteeing the quality of service（QoS）. To solve this problem, we make a trade-off between migration performance and energy consumption. The basic idea of this paper is to concentrate workloads into a minimum set of nodes and close the idle ones to save energy. At the same time, we consider the performance of Virtual Machine migration to guarantee the QoS. The experiment result shows that this strategy can reduce the number of open node and give consideration to migration energy consumption and migration delay. So it guarantees the QoS in the premise of saving energy.

关 键 词：绿色云计算 虚拟机 动态迁移 任务聚合 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]