Research And Implementation Of Virtual Machine Deployment Technology In The Cloud Environment

In recent years, cloud computing technology has got wide attention, and currently it is a research hotspot in academic and industrial. Because of having enhanced cloud computing system flexibility and extensibility, improved the efficiency of resource use, reduced the hardware investment and maintenance costs, and many other advantages,virtualization technology is widely used, and it has become the key technology in cloud computing field.Virtual machine deployment technology is one of the highlights in virtualization field. It mainly focuses on using appropriate strategy and algorithm to deploy virtual machines to physical machines, and make the virtual machines run efficiently on cluster. Existing deployment algorithm mainly aimed at CPU and other single factor, which usually does not take into account the load type of virtual machines, so there are some limitations;In addition, existing virtual machine rapid deployment method based on the image mode has two serious problems including low customer satisfaction and complex management. According to the above problems in the virtualization platform Xen, the finished specific work is described as follows:(1) According to defects of caring single factor and the unclear target in the existing virtual machine deployment strategy and algorithm, two kinds of virtual machine deployment algorithm are proposed, including the virtual machines deployment algorithm based on Weighting matrix (MW-EC) and the virtual machines deployment algorithm based on Genetic Algorithm(GA-LB). The former mainly focuses on quality of service while considering the energy consumption of cluster system and it can effectively improve quality of service; While the latter is to improve the original design on the basis of genetic algorithm, it have a strong ability of load balancing.(2) In combination with the above two kinds of virtual machine deployment algorithm, this paper designs and realizes a virtual machine deployment management framework(VMPF) in a cloud environment, and the framework is made up of the following parts:virtual machines deployment management module, agent deployment module, resource monitoring module and template management module and other modules. Deployment management module can generate the deployment plan according to customer’s request and virtual machine deployment algorithm; Deployment agent module creates a virtual machine and then completes deployment according to deployment solution; Resource monitoring module mainly collects load information of physical machines, and serves for deployment management module;the template management module provides a virtual machine image management function. The framework can provide users a convenient and quick application to deploy virtual machines, and implements automation that users apply for virtual machines; it has the high deployment efficiency.(3) This paper designed and implemented the proposed virtual machine deployment framework and algorithms, and the proposed deployment algorithms based on weighted matrix and genetic algorithm respectively test the performance and are compared with the existing FF algorithm, FFD algorithm and optimizing algorithm. Experimental results show that the virtual machine deployment based on matrix weighted algorithm can achieve good energy saving targets, while the virtual machine deployment algorithm based on genetic algorithm has better ability for load balancing.