Virtual Machine Memory Management Based On Cloud Computing Environment

With the rapid development of the cloud computing and virtualization technologywidely used, the virtual machine memory management is more and more attention byresearchers. Many efficient virtual machine memory management technologies have beendeveloped. However, the existing virtual machine memory management technology focuson the aspects of the regulation of its own memory and the lack of a mechanism of use theidle memory of remote virtual machine. This reduces the memory utilization of the entirecluster of virtual machines in the cloud computing environment and makes the limitedscope of application of the technology of virtual machine memory management. This thesisis to study these issues for the current appears and put forward the virtual machine memorymanagement model DMBMM based on cloud computing environment.The model contains three key technical points, virtual machine memory informationmanagement strategy, virtual machine memory balancing strategy and memory switchingmethod based on network storage. In the virtual machine memory information managementaspects, this thesis uses the double memory information management structure to manageand statistical information about all virtual machine memory utilization. First, the memorymonitoring module in each virtual machine cluster statistics its own cluster virtual machinememory use of information. Then, each memory monitoring module interact informationwith memory management server and complete statistics of all virtual machines memoryinformation. In the virtual machine memory balancing strategy aspects, based on thememory status of each virtual machine cluster which is scarce state or ample state,implementation of the stand-alone balancing strategy and multi-machine balancing strategy.In the memory switching method based on network storage aspects, DMBMM introducesthe network storage technology and utilizes the existing network to use the remote memory,this solve the problem of the existing virtual machine memory management technology cannot use remote memory.This thesis uses two micro-benchmarks test program developed by MyEclipse toobtain test data. Test and verify validity of stand-alone balance strategy and multi-machinebalance strategy combining with development of software&hardware and the actual environmental conditions of the experiment. By comparing and analyzing of commonvirtual machine memory management model, prove the advantage of this model in effectiveutilization of remote memory and improving memory access speed. Meanwhile confirm thefeasibility of the ideas in this thesis from experiment point of view.