Research On Virtual Machine Deployment Method For Optimizing Energy Consumption In Cloud Computing

With the rapid development of cloud computing technologies and its high availability,scalability,security,and pay-as-you-go features,more and more people are starting to use the services provided by cloud computing.In order to ensure service quality and resource reliability,cloud service providers often deploy too many resources in the cloud data center,which leads to a lot of energy consumption.As the energy consumption problem of cloud computing data centers becomes more and more significant,how to reduce the energy consumption of cloud data centers and achieve efficient use of resources while ensuring service quality has become an urgent problem to be solved.Virtualization technology provides an effective solution for the efficient use of data center resources.It allows cloud service providers to run multiple virtual machines on a single server at the same time.Online migration technology can realize the dynamic integration of virtual machines to a small number of servers that satisfy resource requests.The purpose of reducing energy consumption can be achieved by turning off idle servers.Therefore,the deployment of virtual machines based on virtualization technology has become a research focus at home and abroad.Currently,the initial deployment of virtual machines is mainly based on the degree of performance matching.The lack of consideration of the type of virtual machine load makes it impossible to efficiently use server resources and also leads to resource competition.On the other hand,the current dynamic deployment of virtual machines does not take into account the changing trend of server load and cannot meet the dynamically changing cloud computing environment.For the above two issues,the specific work of this paper is as follows:(1)For the initial deployment of virtual machines,a virtual machine allocation method based on load type awareness is proposed.This method aims at energy consumption optimization and load balancing.At the same time,it considers four types of resources:CPU,disk,network bandwidth,and memory requirements.It minimizes the deployment of virtual machines that consume the same type of resources to the same server.Finally,the problem is solved by an improved ant colony algorithm that incorporates a simulated annealing algorithm.Experimental results show that the proposed algorithm effectively reduces energy consumption.(2)For the dynamic deployment of virtual machines,an efficient scheduling method for virtual machines based on load forecasting is proposed.First use the time series model ARMA to predict the change of server load in advance,and determine the migration timing of the virtual machine through the delay mechanism to avoid frequent migration of the virtual machine.Second,use a simulated annealing algorithm to find a suitable destination server for the virtual machine to be placed.Experimental results show that the proposed algorithm can sharply reduce the number of virtual machine migrations and significantly reduce energy consumption.