Research On Scheduling And Migrating Strategy Of Virtual Machine Based On SLA Under Random Workload

With the growing number of cloud computing users,the user market has been gradually subdividing.The cloud computing resources of cheap cloud services which cloud service providers provide for low-end users are generally sold in overbooking mode.Although the resource overbooking can significantly reduce the cost,there will be a great chance to decline the quality of cloud service,and increase the risk of violating SLA agreement when the workload of datacenter virtual machine fluctuating.In order to ensure the quality of cloud services while selling resources in overbooking mode,cloud service providers use virtual machine migration technology to optimize the resources in the datacenter dynamically.With the continuous expansion of cloud computing and business globalization,the workload fluctuations of virtual machine become difficult to find a law.How to ensure the quality of cloud services has become a need to overcome the difficulties in the case of workload fluctuations.In view of the above problems,this paper launches the research work from the following three aspects.(1)On the basis of combing the related concepts and research status of cloud computing and virtualization,this paper expounds the life cycle of virtual machine.In addition,it discusses the quality of cloud computing services in depth and summarizes the factors that affect the quality of service and the quality indicators of cloud computing service.(2)This paper studies the influence of the combination of different virtual machine selection strategies and virtual machine placement strategies in double-threshold virtual machine migration strategy on the quality of cloud computing services under the environment of random load fluctuation.The experimental results show that the doublethreshold virtual machine migration strategy can reduce the number of virtual machine migration and improve the quality of cloud service in the environment where the virtual machine load fluctuates smoothly,but does not perform well in random load fluctuation environment.In addition,using first-fit descending(FFD)strategy as the virtual machine placement strategy,the double-threshold virtual machine migration strategy works outstanding in the resource overbooking environment.For the virtual machine selection strategy,the migration times of virtual machine could be reduced by using highest-growth(HG)strategy and the quality of service could be improved by using largest-requirement(LR)priority strategy.(3)An improved method based on double threshold virtual machine migration strategy is proposed.In this paper,a triple-threshold virtual machine migration strategy is proposed based on the upper threshold,the migration threshold and the lower threshold for the random load fluctuation environment,and the threshold setting of the triplethreshold virtual machine migration strategy is studied.The experimental results show that the triple-threshold virtual machine migration strategy can effectively reduce the number of virtual machine migration and improve the quality of cloud computing service in the random load fluctuation environment compared with the double threshold virtual machine migration strategy.