Researches On Virtual Machine Migrating Algorithm To Save Energy In Cloud Datacenter

In recent years, with the development of cloud computing technology, more andmore resources are concentrated in cloud datacenter, and it brings challenges to themanagement of datacenter energy consumption. The high energy consumption ofdatacenter not only wastes energy and instability of the system, but also has adverseeffects on the environment. Therefore, with the use of cloud datacenter resources, theproblem of high energy consumption should be considered. As an importantfoundation of cloud computing, virtualization technology plays an important role indatacenter resource management. In order to save more energy of the datacenter andhow to use virtualization technology to improve the resources utilization of datacenterand reduce energy consumption become a research hotspot in recent years.This thesis analyzes and researches the energy consumption of cloud datacenterfrom the aspects of power measurement and power modeling, for the difficult problemof measuring power of virtual machines, proposes the monitoring and measurementmethods of power of virtual machines, and creates the power model of virtualmachines and the power model of virtual machine migration. The common migrationmethods use heuristic algorithms to allocation virtual machines and the solution resultsis easy to fall into local optimal solution, based on the study of Genetic Algorithm, animproved algorithm called Migrating algorithm based on Genetic Algorithm (MGA) isintroduced in this paper, which roots from genetic evolution theory to achieve globaloptimal search in the map of virtual machines to target nodes, and improves theobjective function of Genetic Algorithm by setting the resource utilization of virtualmachine and target node as an input factor into the calculation process, and makes thenumber of target nodes and migration times least under the conditions of servicelevel agreements, and reduces the energy consumption of datacenter.Finally, the MGA is simulated in the simulation platform Cloudsim, and the paperuses the models to measure and analysis the energy consumption in the realization processes of MGA. There is a contrast between MGA, Single Threshold(ST) and Double Threshold (DT) through simulation experiments, the results showthat the MGA increases the search speed of the algorithm and can effectively reducemigrations times and the number of host machine used, improves the resourceutilization of datacenter and reduces energy consumption.