Clustering Deployment Algorithm For Energy-Efficient Virtual Machines In The Cloud Environment

With the development of cloud computing technology,cloud service providers can provide users with convenient and reliable on-demand services.However,while cloud computing brings many conveniences to humans,it also faces some challenges.As one of the challenges,virtual machine deployment has been widely concerned by academia and industry.How to efficiently deploy virtual machine clusters,shorten deployment time,reduce data center energy consumption and overall bandwidth requirements has become a current research hotspot.The existing virtual machine cluster deployment methods are mainly clustered based on the communication bandwidth between virtual machines,but the resource similarity between virtual machines is not fully considered.Based on this issue,this paper proposes a method of virtual machine cluster deployment based on resource request similarity(SCVMD).The main contributions of this thesis include:(1)A method of constructing virtual machine clusters based on the similarity of resource requests was proposed to solve the problem of constructing virtual machine clusters.Quantify the similarity between virtual machines by integrating the similarity between various resources to form an virtual machine cluster that interact and communicate with each other;The virtual machine cluster is projected to the plane to form a binary graph interconnected by several vertices;Finally,the weights on each side of the binary graph are determined according to the communication demand matrix between virtual machines to form a weighted undirected graph of virtual machines.(2)A virtual machine cluster deployment method based on graph cutting theory was proposed to solve the problem that a single physical host cannot meet the large scale of virtual machine clusters.The weighted undirected graph of virtual machines is described by energy minimization,and the problem of virtual machine cluster deployment is transformed into the problem of maximum flow and minimum cut;the method of increasing flow and labeling are used to cut the virtual machine cluster graph to ensure the minimum energy consumption after cutting;Finally,the virtual machine cluster after cutting is deployed to the physical host according to the principle of maximum remaining resources.(3)Experiments with the method in this paper on the cloud simulation platform,and compared with the single virtual machine deployment method SVMD,virtual machine cluster deployment methods VMDGT and VMDCT,the simulation results show that the method in this paper is in terms of virtual machine deployment time and energy consumption all aspects have certain advantages.