云环境中基于动态监测周期的动态资源管理模型.pdfVIP

2016 年3 月第3 期 小型微型计算机系统 Journal of Chinese Computer Systems Vol. 37 NO.3 2016 云环境中基于动态监测周期的动态资源管理模型 袁健，陈冬露 (上海理工大学光电信息与计算机工程学院，上海2仪盼到 E-mai1: 归anjianwq@ 163. com 摘 要:为了实现云平台虚拟机的资源负载均衡，提出一种基于动态监测周期的动态资源管理模型.在监测模块采用基于动态 监测周期的自适应资源监测算法，它根据物理服务器的资源负载状态确定动态的监测周期，提出了当结点处于轻度过载或重度 过载时，物理服务器自发地执行过载均衡方法.在选择虚拟机迁移时考虑虚拟机资源的潜在增长率，提出了改进的多因素虚拟 机选择算法.选择目标物理服务器时则采用降序最佳适应算法.实验结果表明，该模型能有效减少虚拟机迁移次数，降低监测机 制耗能，尽可能保证了负载均衡. 关键词:资源管理;云计算;虚拟机;负载均衡 中图分类号:1凹02 文献标tR码:A 文章编号:1αlO-1220 (2016) 03 -0474-05 Resource Management Strategic Model ßased on Dynamic Monitoring Period in Cloud Envi- ronment YUAN Jian ,CHEN Dong-1u (Sch∞，lof 句tical E1ectrical Compu也r Engineering ,University of Shanghai f or Science Techno1ogy ,Shanghai 2αlO93 ,China) Abs仕act: In order to achieve the balance of resoUI四 10ad on the cloud computing p1atform ,a dyn缸咀c reso山ce management strategy ( DRMMDMP) was put forward ,which is based on the d归缸咀c monitoring period techno1ogy. The strategy provided a se1f-adap出g resource monitoring algorithm based on dynamic monitoring period for physical servers ,which was determined by 由e resource 10ading state of physical servers. 刊.e strategy was execu能d， when physical servers were in light-overload or weight-overload. To take the pc• tential grow由 rate into account in the pr田ess of se1ecting virtual machines for migration ,it was proposed that an improved and mu1ti- factor se1ection algorithm for virtual machines. Also. the descending best-fit algorithm was provided in se1ecting target physical server. 刀le resu1ts indicated that the strategy can reduce effective1y the times of migration for virtual machines and energetic consumption for monitoring mechanism ,better 伊缸阻能e the balance of 10ad. Key words: resource management j cloud computing j 悦rtual machine j 1