Model Construction Of Virtual Machine Cluster Deployment Strategy In Cloud Computing

As a commercial resource use and service model,cloud computing has been increasingly applied to all walks of life.Based on virtualization technology,it breaks the coupling relationship between the underlying resources and the upperlevel services,and enables users to use various services on demand at any location and using various types of terminals.However,in the face of a massive task request,with a single virtual machine of the traditional deployment way to perform tasks,the deployment efficiency and resource utilization rate is very low,but through the cluster approach to deploy virtual machines can not only solve the problem,but also can reduce the use amount through the physical host,to achieve the aim of reducing energy consumption.Based on the above,this paper starts from two levels in the deployment of virtual machine cluster.The first level is based on the method of clustering multiple virtual machines,and proposes a multi virtual machine clustering method based on similar performance.In the method,multiple virtual machines with similar performance form an interconnected virtual machine cluster through network connection.The virtual machines in the cluster communicate with each other and project the relationship on the plane to form a graph with multiple vertices interconnected.This method takes full account of the consistency of these virtual machine tasks and the similarity of resource requirements,and provides a favorable premise for the deployment of virtual machine clusters.The second level is to deploy these virtual machine clusters on the basis of virtual machine clusters,and proposes a virtual machine cluster deployment model based on graph segmentation.The deployment process is performed in two situations: one case is: the physical host When the virtual machine cluster's deployment resource conditions can be met,if it also satisfies the pre-deployed physical host's CPU overload threshold,it is selected for deployment;in another case,the number of virtual machines in the cluster is higher than The physical host list that performs the task is queried for physical hosts that meet the virtual machine resource requirements.Many of these do not meet the resource requirements of the virtual cluster.To solve this problem,this paper applies the graph-cutting theory to perform the virtual machine cluster partition and then divides the virtual machine cluster.Sub-cluster in accordance with the first case of the steps deployed to the physical host.Finally,the paper introduced the CloudSim cloud computing simulator in detail,built and extended the cloud platform,and compared the other two classical deployment algorithms as comparison algorithms.The experimental results show that this method has a good effect in improving resource utilization and reducing energy consumption.