蔡立军+何庭钦+孟涛+陈磊

摘 要：为了均衡分布式数据中心物理主机多维资源的利用率，减少物理主机使用数量，节约能耗，提出了一种基于层次拓扑树的虚拟机节能分配算法HTES（hierarchical topology energy saving），此算法可以有效提升虚拟机分配效率.利用Laplacian矩阵，对大规模网络拓扑分割，建立了层次拓扑树模型.基于层次拓扑模型，根据虚拟机请求中IP地址与数据中心的距离，将虚拟机请求分组，从层次拓扑树模型中查詢合适的物理主机区域，按虚拟机请求与物理主机的资源匹配度进行虚拟机的分配.将HTES与其他3种算法进行模拟仿真实验，从虚拟机分配时间、资源均衡率、能耗和物理主机使用情况等方面验证了HTES算法能够有效加快物理主机搜索速度，增加底层占用物理主机的集中度，降低底层物理主机的使用数量，达到节约能耗的目的.

关键词：数据中心；虚拟机分配；层次拓扑树；能源利用率

中图分类号：TP391 文献标志码：A

Energy Saving Allocation Algorithm

of a Virtual Machine Based on Hierarchical Topology Tree

CAI Lijun1，HE Tingqin1，MENG Tao1，CHEN Lei2

（ 1. College of Information Science and Engineering，Hunan University，Changsha 410082，China；

2. College of Electrical and Information Engineering， Hunan University，Changsha 410082，China）

Abstract：In order to balance the utilization of multi-dimensional physical host resources and reduce usage number， a virtual machine allocation energy saving algorithm was proposed based on hierarchical topology tree （HTES） in the environments of distributed data center， which enhances allocation efficiency of the virtual machines. Laplacian matrix was then used to split large-scale network topology and to build hierarchical topology tree model. Furthermore， according to the distance between request IP address and the data center， HTES divided the virtual machines into groups， and searched the appropriate physical host region from hierarchical topology tree for allocation， which is based on the match degree between virtual machine requests and physical hosts. Simulation experiments were performed on HTES algorithms and other three algorithms， considering the virtual machine allocation time， resource balancing rate， energy consumption and physical host usage， and other aspects. The results shows that the HTES is able to balance multi-dimensional resources physical hosts， reduce physical host usage， and save energy consumption.

Key words：data centers；virtual machine allocation；hierarchical topology tree；energy utilization

随着信息科技的不断发展，数据中心作为一种基础设施，已经被各行各业普遍使用.然而当前数据中心的发展也面临新的问题，数据中心的规模不断扩张，地理位置愈趋分散.多个分散的数据中心通过高速网络互联，共同组成了大型的分布式数据中心.在分布式数据中心内，用户通过按需付费的模式，向数据中心提交需求.数据中心根据用户地理位置，从较近的基础设施库分配资源并构建虚拟机为用户服务.然而，大规模分布式数据中心环……