Research On Service Migration Based On Edge Computing In Internet Of Things

In recent years,mobile edge computing(MEC)has developed rapidly and has become one of the key technologies to realize the vision of the Internet of Things.Sinking computing and storage capabilities from the cloud center to the edge of the network helps to get rid of the restriction of transmission delay and network bandwidth.However,due to the user mobility and limited resources of edge nodes,some edge nodes cannot provide users high-quality services.In view of this,this thesis studies the service migration strategy based on MEC,and migrates some services from the original node to other nodes that meet the requirements to improve the quality of service.This thesis designs service migration strategies under two modes: centralized mode and distributed mode.In the centralized mode,the edge server reports all information to the network controller,and the network controller decides which service to migrate to which server based on the received information and collected information.Assuming that the storage resources in the server are sufficient,and the hotspot information will be allocated to each server by the local network controller and periodically updated.Therefore,in the centralized mode,this thesis mainly aims at computing-intensive services.By establishing a service migration model with the goal of minimizing delay and energy consumption,the service migration is mapped to a nonlinear 0-1 programming problem.In order to solve this problem,this thesis designs a particle swarm based service migration mechanism(PSSM),which includes queue delay prediction strategy(QDP),delay-aware computing resource allocation strategy(DCRA)and quantum particle swarm based service migration strategy(QPSM).Simulation results show that the proposed scheme in centralized mode not only effectively reduce service delay and energy consumption,but also greatly improve the processing efficiency of the server.In the distributed mode scenario,there is no local network controller,and its functions are delegated to the servers belonged to base stations.The cache information cannot be reasonably distributed to each server through the local network controller,so the initial cache placement and cache update need to be considered.In the actual environment,many servers in the base station belong to different operators,and some private information will not be disclosed.Each server will make decisions to maximize its own interests and will not make joint decisions with other servers.Each server makes independent decisions based on its own status,benefits and the estimation for the remaining servers.Therefore,this thesis designs a service migration scheme based on Bayesian game,including initial cache placement algorithm,cache update algorithm,and Bayesian game based service migration algorithm.In order to verify the effectiveness of the proposed scheme,this thesis simulates and analyzes the changes in the average server revenue,average user payment price and average winning rate under different server numbers.Moreover,the simulation verifies that the proposed migration scheme can effectively reduce the user's service delay.