基于混合粒子群算法的虚拟数据中心能耗优化  

Energy Consumption Optimization in Virtualized Data Center Based on Hybrid Particle Swarm Algorithm

作　　者：敬思远[1] 佘堃[1] 

机构地区：[1]电子科技大学计算机科学与工程学院,成都611731

出　　处：《计算机工程》2012年第15期276-278,282,共4页Computer Engineering

基　　金：国家"863"计划基金资助重大专项(2008AA04A107);国家科技支撑计划基金资助项目(2009BAH46B0302)

摘　　要：针对当前数据中心节能整合研究中仅考虑服务器能耗的现状,提出一种同时考虑服务器和网络设备能耗的方法。该方法通过感知数据中心的网络拓扑,使运行的服务器和网络设备最少,以此最小化能耗。对问题进行新的形式化建模,并设计一种混合的粒子群优化算法HPSO-NA来实现虚拟机整合。实验结果表明,该方法能有效降低整体能耗。Since current researches on energy efficient consolation in data center only consider part of energy consumed by servers,this paper proposes a novel method considering both servers and network devices.The proposal achieves energy minimization by minimizing the active servers and network devices through analyzing the net topology.Moreover,the problem is modeled and a hybrid particle swarm algorithm called HPSO-NA is designed to implement virtual machine consolidation.Experimental results show that the proposed method can effectively reduce the energy consumption.

关 键 词：能耗最小 网络感知 整合 虚拟机迁移 粒子群优化 装箱问题 

分 类 号：TP302.7[自动化与计算机技术—计算机系统结构]