Research On Virtual Machine Resource Scheduling Based On Multi-objective Optimization

As a new type of green computing model,cloud computing has changed the traditional service mode to a great extent.It provides users with various application services through close integration with network technology.However,due to the continuous expansion of the cloud data center infrastructure,the cloud computing industry has brought high energy consumption problems.For the dynamic configuration of resources in the cloud computing environment,it is also necessary to consider the stability of resource distribution in various states.Therefore,researching how to optimize the virtual machine scheduling strategy is of great significance for improving the resource utilization of the cloud data center,reducing energy consumption,and improving the stability of physical nodes in the cloud data center.This paper studies the virtual machine scheduling algorithm based on multi-objective optimization.The main contents are as follows:(1)Considering the problem of high energy consumption generated by cloud data center and how to improve the utilization rate of cloud data center resources during the initial stage of virtual machine placement,a virtual machine initialization placement algorithm(M-GA)based on energy optimization is proposed.Based on the genetic algorithm,the algorithm improves the search efficiency by designing the fitness function,establishing the server energy consumption model and the multi-dimensional resource model,by defining the genetic evaluation parameters,and optimizing in terms of crossover and mutation.The results of CloudSim simulation experiments show that M-GA is better at improving data center resource utilization than traditional genetic algorithm(GA).At the same time,M-GA can reduce data center energy consumption.(2)In the dynamic management phase of virtual machine,the virtual machine resource scheduling algorithm(M-MOGA)based on multi-objective optimization is proposed for the dynamic change of data center load.The algorithm considers the cost of virtual machine data migration,the stability of the new virtual machine distribution state,and the service level agreement(SLA)in the new and old state transition process.The results of CloudSimsimulation show that compared with the genetic algorithm(GA-S)for distributed stability of multiple virtual machines,M-MOGA can better balance the stability of multi-virtual machine distribution,migration cost of virtual machine relocation,and old and new.The relationship between the three service level agreements(SLA)in the state transition process.