Research On Many-objective Virtual Machine Placement And Migration

Cloud computing,as a new computing model and business model,allocates tasks to resource pools made up of a large number of computers,enabling users to acquire computing power,storage space,and application services on demand.In recent years,with the development of cloud computing,more and more users have deployed their own applications to cloud data centers.These applications have different resource requirements,so that the loads of the virtual machines are in constant changes.Faced with the diversified requirements of users,how to allocate suitable physical machines for virtual machines,and how to migrate virtual machine on physical machines with abnormal load under random load fluctuations of virtual machines have become a concern.In view of the above issues,this paper carries out research work from the following two aspects.(1)For the virtual machine placement problem,a many-objective virtual machine placement optimization model is built.Based on the actual problem scenario,an improved EEKnEA algorithm is proposed to solve the model.First,in order to meet the diverse scheduling requirements,a virtual machine scheduling model that considers load balancing,resource utilization,energy consumption,and robustness indicators is constructed.The KnEA algorithm which has better effect to solve the many-objective optimization problem is improved.The heuristic rule that the virtual machines with frequent communication and heavy traffic are placed on the same physical machine as much as possible is introduced,and the above model is optimized and solved.The experimental results show that compared with the existing algorithms,the proposed algorithm can not only obtain a better quality solution set,but also achieve the balance between multiple objectives.(2)For the problem of virtual machines migration when physical machines have abnormal load under random load fluctuations of virtual machines.A many-objective virtual machine migration optimization model is proposed and an improved KnEA-based algorithm is proposed.In addition to considering the impact of virtual machine migration on load balancing,resource utilization,energy consumption,and robustness indicators of allocation schemes,the migration cost for virtual machines was also be considered.For the optimization algorithm,two heuristic rules are added in the population initialization to improve the solution efficiency while obtaining a better initial population.The many-objective virtual machine placement and migration research proposed in this paper can further meet the diverse requirements of users for resource requests,and at the same time ensure the low energy consumption and high resource utilization of cloud data centers,so that cloud data centers can provide more stable services.Thus,it has good practical application prospect.