Research On The Batch Deployment And Dynamic Migration Of Virtual Machine In Cloud Computing Environment

Cloud computing has so many advantages as distribution of demand, low cost, high reliability and good scalability, which has developed into the focus of academic world and the industrial world in a relative short period of time. The long-term development of the key technologies such as virtualization, mass storage and programming model, have created favorable conditions for the rapid development of cloud computing. With the rapid development of cloud computing technology, the resource scheduling and management of data centers becomes more and more important. However, the current data center tend to have the problems of unbalanced load of physical machine and the low resource utilization; Moreover, with the dynamic changes of the tasks in the data center, the problem of high energy consumption should not be ignored, which calls for the dynamic migration of virtual machine to reduce its energy cost.At present, most researchers consider the virtual machine deployment as a 0-1 knapsack problem and use the intelligent optimization algorithms to solve, but all kinds of intelligent optimization algorithms have their own advantages and disadvantages; At the same time, the most existing research simplify the processing of resources in data center,which leads to the shortcomings of research on the dimension of resource. In view of the current status of research, this paper takes the minimization of the imbalance load degree of the data center as the optimization target, and taking four dimensions of resources of the data center into account; Improving the particle swarm algorithm, which only has significant effect on the continuous problem; And combining the simulated annealing algorithm, so that it is able to solve the problem of discrete virtual machine deployment. Experimental results shows that the virtual machine batch deployment algorithm,which based on discrete particle swarm optimization algorithm, has obvious advantages.At present, the algorithms of selecting physical machine for the virtual machines in migration just simply consider the reduction of energy cost of data center as the optimization goal, and ingore the influence of the load of physical machine on the migration times of virtual machine. In this paper, we put forward the algorithm for selecting physical machine based on the performance label and the awareness of energy, taking the energy cost of data center and the load levels of physical machine into account. Experiments show that the algorithm presented can effectively reduce the energy consumption of data centers and the times of virtual machine migration.