Research On Virtual Machine Resource Scheduling In Cloud Data Center

With the rapid development and popularization of cloud computing,the problem of high energy consumption,high cost and low efficiency of cloud data center is highlighted increasingly.In this context,the key requirement on future data center is to optimize the resource management method and energy consumption mode of data center while guaranteeing the reliability of cloud computing,which is also the precondition of ensuring sustainable development of data center in the future.Benefiting from technological advance and hardware support,the virtual technology springs up again.It takes virtual machine as particle to implement fast and new deployment of resources,thereby enhancing resource usage.Nowadays dispatching management of virtual machine resources has become a hot research topic.By learning the related technology of virtual machine resource scheduling,this paper analyzes the problems of high energy consumption,low efficiency and unbalanced load existing in cloud data center,and launches the research on the placement and migration of virtual machine resources in cloud data center.The main research work of this paper is as follows:Firstly,Aiming at problems of energy consumption and imbalance existed in server's internal source load of cloud computing center,a virtual machine placement algorithm has been presented,which could optimize the energy-efficiency of cloud data center.Based on the grouping genetic algorithm,the energy consumption model and load balance measurement index are established.Meanwhile,a multi-dimensional fitness function is constructed on the foundation of standard deviation and resource utilization,which could finally guide the solution searching.Furthermore,the optimized crossover and mutation are provided to improve solution quality.Secondly,As virtual machines with static threshold cannot optimize energy consumption and load balancing of system according to cloud loads in real time,an optimized virtual machine migration algorithm is proposed.This is a method to dynamically change threshold range by forecasting task volume of data center,then select the proper time,migration object and target host and combine migration strategy to complete dynamic migration and integration of virtual machine.This method can reduce the workload of servers so as to minimize the energy consumption of data center.Finally,we extend the scheduling algorithm based on CloudSim simulation platform,and compare the performance of the algorithm with other scheduling algorithms.Experiment showed that the scheduling algorithm proposed in this paper is feasible and effective.