Research On Dynamic Virtual Machine Scheduling Strategy Based On Improved Genetic Algorithm

Cloud computing has been widely used in education,government,finance and other fields,and has opened up new ways of using Internet resources.OpenStack is one of the most popular open source IaaS cloud platforms,but OpenStack still has many shortcomings in virtual machine scheduling,resulting in unbalanced hardware usage of physical hosts,resulting in degraded virtual machine performance.If the hardware resources of the physical host can be reasonably utilized,the performance of OpenStack will be greatly improved.The paper focuses on the filter policy selected by default in the original virtual machine scheduling policy of OpenStack.Firstly,a dynamic virtual machine scheduling model is built according to the dynamic virtual machine scheduling problem,and a genetic algorithm is introduced into the scheduling strategy.The fitness function of the genetic algorithm comprehensively considers the three factors of CPU utilization balance,memory utilization balance and virtual machine dynamic migration cost of the physical host cluster in the data center.Secondly,aiming at the problem of low efficiency and slow convergence of genetic algorithm and poor readability of the chromosome,the coding method,initial population,cross operation and mutation operation of the genetic algorithm are improved.Tree coding is designed to improve code readability and computational efficiency,cross operation and mutation operation are designed for tree coding.Prototype-like chromosomes are proposed during the initial population phase,which produced a better initial population and shortened the convergence time of the algorithm.Finally,a hybrid strategy based on improved genetic algorithm is proposed.The improved genetic algorithm and filter strategy are combined to make the improved genetic algorithm and filter strategy run alternately,and the dynamic virtual machine scheduling efficiency is improved.In order to prove the effectiveness of the proposed dynamic virtual machine scheduling strategy based on improved genetic algorithm,the CloudSim simulation platform is used for verification.The experimental data proves that the proposed hybrid strategy makes the resource utilization of each host more balanced,effectively solving the problem of load imbalance and large dynamic migration overhead in virtual machine scheduling.Moreover,experiments have shown that the improved genetic algorithm has a faster convergence speed and a higher optimal solution quality.