Research And Implementation Of Virtual Machine Migration Algorithm In Cloud Environment

Virtualization technology,as a key technology of cloud computing,enables multiple virtual machines to run on a single host to meet the operational needs and environmental requirements of different applications and improves the efficiency of the host.However,the resources of the host are limited.When the virtual machine runs too many tasks,the host will be overloaded and exception occur.In view of the above problems,considering the communication cost of virtual machine migration,the research content and main contributions of this paper are as follows:(1)The traffic-aware virtual machine migration based on double auction is studied.This study considers the communication between the virtual machines inside the host and establishes the communication mode of the data center.The virtual machine migration algorithm is mainly divided into two parts.First,the virtual machine to be migrated is selected on the overloaded host according to the communication and the occupied resource size of the virtual machine,and then the virtual machine to be migrated is allocated an underutilized host with the goal of minimizing the communication cost.In the first part,this paper proposes a virtual machine greedy selection algorithm(VMs-GSA)to select the virtual machine to be migrated on the overloaded host.In the second part,this paper proposes a novel virtual machine migration algorithm based on bilateral auction(VMM-DAM).VMM-DAM motivates buyers(virtual machines)to match sellers by allowing sellers(low-load hosts)to offer discounts,with the goal of minimizing communication costs,thereby obtaining an allocation between the virtual machine to be migrated and the low-load host.Finally,the paper proves the economic attributes of the auction mechanism theoretically.Experiments show that compared with the communication cost generated by random algorithl,the communication cost generated by VMs-GSA is reduced by 35%.VMM-DAM can solve host overload problems while reducing communication costs.(2)The traffic-aware virtual machine migration based on gene aggregation-based genetic algorithm is studied.This study not only considers the communication between the internal virtual machines of the host,but also considers the communication between the hosts and establishes a comprehensive communication model.A novel genetic coding method for gene aggregation was proposed.The coding method considers communication between virtual machines for gene aggregation operations,effectively reducing the number of genes of chromosomes.After using VMs-GSA to select the virtual machine to be migrated on the overloaded host,the constrained gene aggregation genetic algorithm is used to find the mapping relationship between the virtual machine to be migrated and the underutilized host that minimizes the communication cost,and finally complete the virtual machine migration.Experiments show that virtual machine migration based on gene aggregation-based genetic algorithm reduces search time and communication cost compared with traditional genetic algorithm-based virtual machine migration.(3)A prototype system for virtual machine migration algorithm simulation was designed and implemented.Users need to register first and then log in to enter the system.The system can use the virtual machine and host parameter information input by the user to create a simulated data center,and realize the virtual machine migration function based on double auction-based and genetic aggregation-based genetic algorithm.Finally,prototype system gives an evaluation method for virtual machine migration algorithm.In summary,this paper proposes a set of virtual machine migration algorithms in cloud environment.Experiments show that its effectiveness can be applied to the actual scenario of virtual machine migration.