An Improved Virtual Machine Migration Strategy In Edge Computing

With the further development of the Internet and artificial intelligence,emerging applications such as virtual reality,telemedicine,and autonomous driving have strict requirements on latency.However,in the traditional cloud computing model,the application that provides services to users is usually in a virtual machine on a cloud computing center far from the user,and the communication delay is high;with the explosive growth of smart terminal devices in recent years,the data generated by these devices has also exploded,if all these data are uploaded to the cloud computing center for processing,it may cause network congestion near the cloud computing center,resulting in higher communication delays.And the linearly increasing computing power of the cloud computing center cannot match the exponentially increasing amount of data.Edge computing adds computing resources between the user device and the cloud computing center,so that user data can be partly calculated and stored in the path uploaded to the cloud computing center,reducing the network burden and ensuring the user experience.However,some users in the edge computing scenario are mobile.When the user is away from the edge computing node where the virtual machine serving him is located,the service request response delay will become high,which will affect the user experience.If the virtual machine that provides services to users is migrated to an edge computing node near the user,although the service request response delay will be reduced,the service may be temporarily interrupted during the virtual machine migration process,and the migration cost will also be generated.And these will affect the user experience.Therefore,an appropriate strategy is needed to decide whether,when and where to migrate a virtual machine.In order to address these problems,the paper mainly studies the virtual machine migration strategy in the edge computing environment.The paper proposes a virtual machine migration strategy based on multi-attribute decisions that considers computing power,bandwidth,network communication cost,and energy consumption.Aiming at user mobility in edge computing scenarios,the paper proposes a method for estimating network communication costs using a future distance model,and experimentally determines the best degree to consider future distance forward;in multi-attribute decision-making,the weight of each attribute is calculated using the entropy weight method,and TOPSIS(Technique for Order Preference by Similarity to an Ideal Solution)is selected through experiments to calculate the comprehensive evaluation value of each scheme.Based on this strategy,the entire process of virtual machine migration is established,including the trigger time of the migration strategy,the range of candidate migration nodes,and the decisionmaking process.Finally,the simulation experiments of the virtual machine migration strategy proposed in this article and two typical virtual machine migration strategies are compared.The network environment considers randomly distributed network environment and a cellular network environment;user motion trajectories consider linear motion trajectories,L-shaped motion trajectories,and circular motion trajectories,respectively;the load situation takes into account the low,medium,and high load of computing power,and the low,medium,and high load of bandwidth.The performance of different virtual machine migration strategies is analyzed in six evaluation indicators: computing power,bandwidth,unit energy consumption,communication cost,number of migrations and migration cost.The results show that the performance of the proposed virtual machine migration strategy is better than the other two typical virtual machine migration strategies,which effectively reduces communication cost,number of migrations and migration cost.