Research On Virtual Machine Deployment Strategy Of Cloud Computing System Considering DVFS

In recent years,with the rapid development of Internet technology,cloud computing,because of its flexible service model and on-demand business model,accesses to a wide range of applications and development.Cloud data center has been expanding,the data center of the high energy consumption problem has become increasingly prominent.How to improve the efficiency of cloud data center resource utilization,reduce cloud data center power consumption,reduce cloud service providers' rising electricity costs and ease the high energy consumption caused by environmental pollution problems,cloud data center sustainable development is unresolved problem.The main means of energy saving in the data center are divided into dynamic voltage and frequency scaling(DVFS)of the server layer and a reasonable resource allocation strategy of the cluster layer.Virtualization technology can abstract the coarse-grained physical resources into fine-grained virtual resources and flexibly integrate resources to improve the resource utilization of cloud data centers.Virtual machine deployment is a virtual server deployed to the cloud data center of the physical server and a cloud computing resource allocation strategy,that its essence is a NP-Hard combination optimization problem.In recent years,DVFS and virtual machine deployment strategy has become an important means of improving resource utilization and reducing data center energy consumption in cloud computing,and has gained extensive attention in academia.Based on the in-depth study in the DVFS technology,virtual machine migration technology and virtual machine deployment strategy,this thesis proposes an energy model based on DVFS technology,and introduces the calculation method of virtual machine migration overhead in the virtual machine deployment,proposes a virtual machine deployment strategy,and the introduction of multi-core virtual machine model in which and put forward a new virtual machine deployment algorithm.The main contents of this thesis is as follows:1.on the basis of in-depth study of the physical server architecture and DVFS technology,a energy consumption computing model based on multi-core physical servers and CPU execution frequency is proposed.2.on the basis of in-depth study of the virtual machine pre copy migration technology,a method is proposed for calculating the migration cost considering the energy consumption of virtual machine migration,migration time and downtime.3.based on extensive research on virtual machine deployment problems,put forward a virtual machine deployment optimization model,taking into account the cloud data center energy and virtual machine migration overhead,and on this basis put forward a virtual machine deployment algorithm EEVD(Energy Efficient VMs Deployment).4.in view of the existing virtual machine deployment algorithm,without considering the multi core attributes of virtual machine,this paper introduces the multi-core virtual machine,and analyzes the necessity of deployment constraints of multi-core virtual machine model and multi kernel virtual machine.On this basis,a multi-core virtual machine deployment algorithm,EMCVD(Efficient Multi-Core VMs Deployment),is proposed,and the effectiveness of the algorithm is verified by simulation experiments.