云数据中心基于异构工作负载的负载均衡调度方案  

A Load-balancing Scheduling Scheme Based on Heterogeneous Workloads in Cloud Data Centers

作　　者：黎红友[1] 杜吉成 彭舰[1] 王江勇[1] 

机构地区：[1]四川大学计算机学院,四川成都610065

出　　处：《四川大学学报（工程科学版）》2013年第S1期112-117,共6页Journal of Sichuan University (Engineering Science Edition)

基　　金：四川省苗子工程资助项目(2012ZZ036);网络与交换技术国家重点实验室基金资助项目(SKLNST-2010-1-03)

摘　　要：针对目前云数据中心的负载均衡调度方案没有研究将不同工作负载的应用程序进行整合的问题,提出了基于异构工作负载的静态和动态负载均衡调度算法。当放置新的虚拟机的时候,通过静态负载均衡调度算法,将虚拟机分配给物理机。云数据中心运行过程中,采用动态负载均衡调度算法将负载高的物理机上的虚拟机迁移到负载低的物理机上。仿真实验结果表明,新的调度方案不仅能够降低云数据中心的不均衡率,而且使节点内的多维度资源得到了合理的利用。In order to treat the problem that the current approaches to improve the rate of load balance in cloud data centers does not investigate the problem of combining different workload types,the static and dynamic load balancing scheduling algorithms based on heterogeneous workloads were proposed.The new VM was allocated to a certain PM using the static load balancing scheduling algorithm.And the VMs on high load PM were migrated to the low load PMs using the dynamic load balancing scheduling algorithm.The simulation results showed that the new load balancing scheduling scheme improves the rate of load balance in cloud data centers and the multi-dimensional resources in a single node has good balanced utilizations.

关 键 词：负载均衡 调度 云数据中心 虚拟机 

分 类 号：TP308[自动化与计算机技术—计算机系统结构]