虚拟化数据中心的负载分布优化策略  

Load distribution optimization strategy of virtualized data center

作　　者：魏秀然[1] 惠向晖[1] WEI Xiu-ran;HUI Xiang-hui(College of Information and Management Science,Henan Agricultural University,Zhengzhou 450046,China)

机构地区：[1]河南农业大学信息与管理科学学院,河南郑州450046

出　　处：《计算机工程与设计》2022年第8期2109-2114,共6页Computer Engineering and Design

基　　金：河南省重点科技攻关基金项目(152102210112);河南省教育厅科学技术研究重点基金项目(13A520713);2017年国家大学生创新创业训练基金项目(201710466046)。

摘　　要：在虚拟化数据中心中,分布式应用的性能受到资源供给和通信效率双重影响。针对这一问题,将通信开销及资源利用均衡性同时作为优化对象,设计面向虚拟化数据中心的容器分布优化系统Themis。提出容器的主导需求衡量方法,基于此设计通信感知的降序匹配算法用于新容器的快速部署。实验结果表明,与现有方案相比,该系统可以更加高效、公平、普适地优化负载均衡和通信开销,提高性能的同时也验证了其科学性。In a virtualized data center,the performance of distributed applications is affected by both resource supply and communication efficiency.In response to this problem,the communication overhead and resource utilization balance were considered as optimization objects at the same time,and the Themis,a container distribution optimization system for virtualized data centers,was designed.The dominant demand measurement method for containers was proposed,and based on this,a communication-aware descending matching algorithm was designed for rapid deployment of new containers.Experimental results show that,compared with the existing scheme,the system can optimize load balancing and communication overhead more efficiently,fairly and universally,improving performance and verifying its scientificity.

关 键 词：虚拟化数据中心 通信开销 负载分布 匹配算法 通信感知 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]