Research On Mobility Management Of User Tasks In MEC System

With the continuous development of the Internet of things and new application scenarios,the future network should meet the needs of ultra-low latency,high reliability,and massive connectivity.Mobile edge computing(MEC)technology can provide a high-performance,low-latency and high-bandwidth network environment by providing cloud computing and storage resources at the edge of the mobile network,enabling users to enjoy an uninterrupted high-quality network experience.In an edge computing network where MEC servers are highly densely deployed,users can offload compute-intensive or latency-sensitive computing tasks to their associated MEC servers.The high overlap of multiple base station coverage areas near the user and the mobility of the user may cause many problems of task uninstallation,such as the selection of the task uninstallation target MEC server,virtual machine migration target MEC server and VM migration decisions.Based on this,this thesis focuses on the mobility management of tasks in the MEC system.The main works are as follows:1.A MEC server selection scheme based on multiple indicators is designed.The scheme comprehensively considers multiple indicators such as uplink and downlink transmission delay,processing delay,energy consumption,VM migration cost,energy efficiency of task offloading and task processing cost.The candidate MEC servers are sorted based on the linear weighting of each parameter,and the MEC server with the best comprehensive performance is selected to provide services for users.The simulation results show that the proposed scheme can reduce the total delay of task processing and meet multiple performance indicators under the condition of guaranteeing users task unloading cost budget and energy constraint.2.A VM migration decision-making scheme based on minimizing task offloading cost is designed.In order to solve the VM migration decision-making problem when users move,the scheme considers that the complex relationship between subtasks is neglected in the existing research.Firstly,a nonlinear mesh task partitioning model is established.Then,based on this model,a minimal task unload cost optimization problem with execution delay constraint is given.Finally,the VM migration decision-making results for each sub-task are obtained by genetic algorithm.The simulation results show that the scheme can effectively reduce the total task offloading cost of users while satisfying the task unloading delay constraint.