摘要
虚拟化与云计算技术的结合,帮助云计算供应商构建了更便捷、可靠以及规模更大的新型数据中心.但当服务器数量不断增加时,一些服务器可能会出现低负载的情况,会对资源及能耗造成大量浪费.服务器整合方法可以将多台虚拟机整合到同一台服务器上,关闭空闲的服务器,达到节能目的.但是传统方法选择待整合服务器时,仅将服务器的状态作为选择依据,未考虑部署在上面的虚拟机或者服务的状态.在传统方法的基础上,在选择待整合服务器时考虑了多个关键因素,如虚拟机状态、服务器资源占用率、服务性能等.利用多目标决策的方法选择出最需要整合的服务器,保证整合过程中的总能源消耗及迁移代价最小,并通过实验验证了本文提出方法的合理性和可行性.
It helps cloud suppliers structure more convenient,stably and large-scale data center because of combining virtualization and cloud computing.When adding numbers of servers,some servers can appear lowload.It wastes larger resources.The method of consolidating servers is that many VM s are concentrated on a server,other empty servers are closed.The method can reduce resources.When traditional methods select servers,the lever of servers is as standards.Its veracity is low.This paper proposes that selecting consolidating servers needs to think about many key factors,for example the condition of VM,resource utilization rate of servers,service performance and so on.Using multi-objective method selects appropriate server.It can make resources consume and migration cost lowest.Experimental results showthe feasibility and effectiveness of the method.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第4期699-704,共6页
Journal of Chinese Computer Systems
基金
中央高校基本科研业务费专项资金项目(N120804001
N120204003)资助
国家自然科学基金项目(61300019)资助
关键词
服务器整合
节能
多目标决策
云计算
server consolidation
energy conservation
multi-objective decision
cloud computing
