Research On Virtual Machine Deployment Algorithm In Cloud Environment

Virtual machine deployment strategy in cloud data center has become an important research topic.At the same time,it become a hot and difficult problem in the field of current research.There equipped a large number of physical servers,which have a large number of different amount computing resources,in cloud computing center.If deploy virtual machine to the servers without any analysis randomly,will result in the performance of load balance of cloud computing system is poor,and the virtual machine execution efficiency is low,but also a large number of physical resources to waste.This paper first puts forward a kind of virtual machine deployment strategy,which based on Bayes classification combined with the first fit decreasing optimization.This strategy algorithm based on the virtual machine needs,and the remainder of the physical server resources is analyzed,and standardized by the Bayes formula can achieve long-term load balancing threshold to find physical server for deployment of virtual machine.The demand of the virtual machine to adapt for the first time after descending sort algorithm to deploy on a physical server not only reduce the number of failed virtual machine deployment,but also reduce launched open unnecessary physical servers to reduce the energy consumption of the whole system.The simulation experimental results show that the new task deployment strategy not only better realize the load balance for a long time,shorten the task completion time,and in the aspect of reducing energy consumption and task deployment failure rate has obvious effect.This paper first put forward by using naive Bayes formula further classified all virtual machine work,virtual machine on the basis of first come,first service deployment.First based on the virtual machine needs,and the remainder of the physical server resources were analyzed,and screened by the Bayes formula has a high performance cluster servers,then the server cluster as a resource pool,statistics needed by the naive Bayes probability when the subsequent allocation of virtual machine again,depending on the naive Bayes classification strategy can quickly find the server can be deployed virtual machine,the virtual machine in the form of first come,first service after classifying algorithm is deployed to the server,both can achieve the effect of load balance can also make the server in the server load balance,but also improve the execution efficiency of the whole system.Simulation experiment results show that the task deployment strategy not only well realize the long-term load balancing in the system,shorten the task completion time,and to reduce the task deployment failure rate aspect also has the obvious effect.