Research On Virtual Machine Deployment Method Based On Optimizing Genetic Algorithm

The problem of virtual machine deployment in cloud computing is to select physical nodes according to virtual machine resources and constraint requirements,so as to effectively reduce the energy consumption of physical machines and avoid waste of physical machine resources.Although cloud computing provides users with unprecedented convenience,how to find a cheap and efficient way to deal with the defects of its resource consumption and high cost has become an important issue of current academic attention.Based on the research of virtual machine deployment methods,the optimization of physical resource utilization and the minimum number of migrations are proposed.The optimized genetic algorithm(OGA)based on improved genetic algorithm is proposed.By redesigning the mutation and mating process in the genetic algorithm,OGA increases the probability of inheriting the excellent genes to the offspring,and achieves the purpose of quickly solving the virtual machine deployment problem.The specific work is as follows:(1)To improve the existing virtual machine deployment strategies and algorithms with a single consideration and ambiguous goals,a multi-objective optimization scheme with minimal resource utilization and minimal migration times was proposed,which effectively improved the efficiency of virtual machine deployment..In addition,OGA is designed based on the original genetic algorithm and has strong load balancing capabilities.(2)Combined with the above virtual machine deployment algorithm,the virtual machine deployment sequence diagram in cloud environment is designed.The sequence diagram framework consists of five modules: user,template management server,deployment management server,monitoring server and virtual server pool.It can be divided into five steps: service request,resource information acquisition,target server selection,copy virtual machine image,and delivery user usage.The framework enables users to quickly and easily apply for virtual machine deployment,and enables users to apply for virtual machine automation,which is highly efficient.(3)Optimize the selection operator,crossover operator and mutation operator to avoid the algorithm completing the search too early,and can not reach the global search,thus affecting the quality of the final solution.In order to make the adaptability of the algorithm change with the evolution process,this paper uses thedynamic change of the cross probability parameter and the mutation probability parameter for formal expression.The performance test of the virtual machine deployment process simulation experiment was carried out on the simulation platform C1oudSim3.0,and compared with the existing random placement algorithm(Random),Vm AllocationSimple algorithm(VAS)and the first adaptation algorithm(FF).Simulation results show that OGA can effectively reduce the number of virtual machine migrations,reduce physical resource consumption,and save the service provider maintenance cost while ensuring the quality of customer service.