Research On Dynamic Migration Strategy Of Virtual Machine In Cloud Computing

In recent years,with the rapid development of mobile Internet,the demand for computing resources is increasing,and the energy consumption of servers is also increasing.With the continuous expansion of scale,the management cost of computing resources is also increasing.Cloud computing model is proposed based on the development of distributed computing,grid computing and parallel computing models.Cloud computing integrates distributed resources to form a virtual huge computing resource pool,which provides resources to users on demand in the form of services,so as to improve the utilization rate of resources,reduce the service independent energy consumption,and effectively alleviate the resource error in the service process.As the core of cloud computing,virtualization technology has been widely concerned.Because of its powerful management mechanism is an effective way to improve the overall resource utilization of the system.More importantly,the virtual machine dynamic migration technology has received a lot of research as the focus of virtualization technology.However,with the rapid growth of user scale and business demand,the data center continues to add new servers,resulting in more and more energy consumption in the data center.At the same time,the resource scheduling of virtual machine cluster is also becoming more and more important with the number of virtual machines in the data center increasing.In addition,the existing research on dynamic migration rarely takes into account the two aspects of data center migration times and load balancing angle.This thesis proposes a dynamic migration strategy of virtual machine in cloud computing environment by considering three aspects of migration time,migration target and migration destination,which can reduce the migration times of data center and ensure load balance.The main work of this strategy is as follows:1)A trigger strategy of virtual machine migration based on multi resources is designed.The virtual machine migration mechanism will start to ensure the quality of service in the data center when resources are overloaded.2)A virtual machine selection strategy based on the maximum evaluation value is designed,which both considers the resource usage state of virtual machine and current server,combines CPU and memory resources,to greatly reduce the migration times of data center,reduce energy consumption,and ensures the load balance of data center.3)A target server selection algorithm based on TOPSIS multi-criteria decision is designed.The server with the least load is selected as the target server by calculating the evaluation value of the server to ensure the load balance between servers.In addition,this thesis analyzes the defects of the existing migration strategy of OpenStack.The existing migration strategy of OpenStack cannot trigger the migration automatically,and the virtual machine to be migrated needs to be specified manually.Moreover,the existing migration strategy has a single algorithm for selecting the target server,which usually uses the server with the largest memory surplus as the migration server.In view of the shortcomings of OpenStack,this thesis proposes a resource scheduling strategy to optimize the dynamic migration of OpenStack.This thesis also tests and analyzes the proposed resource scheduling strategy on CloudSim through a large number of experiments.The experimental results show that the virtual machine selection strategy based on the maximum evaluation value and the target server selection strategy based on TOPSIS proposed in this thesis can effectively reduce the migration times of the data center,and have good performance in green energy saving.