Research On The Deployment Method Of Delay-Sensitive Service For Edge Computing

The mobile Internet network system is developing rapidly.New applications program has higher requirements towards real-time performance,the traditional information technology(IT)architecture is facing great challenges.With the launch and implement of the fifth-generation Mobile Communication Technology(5G),the data transmission capability of the edge computing framework has been progressively enhanced.To upgrade task processing speed and enable user to enjoy better service experience,user could offload tasks to edge computing nodes for processing to get more data and resources in edge computing.Compared to the traditional cloud computing architecture,the edge computing architecture has lesser network delay,which saves the resources and power consumption of user terminal device.The prerequisite of edge computing to process task requests instantly is computing nodes deploy services corresponding to user tasks.Consequently,service placement strategy in edge computing is a major impact factor of the performance of latency-sensitive applications.According to the attributes of delay-sensitive applications and different application scenarios,this paper studies and explores service placement strategies in the edge computing framework.1.A service placement strategy for joint network selection and resource scheduling in edge computing.The service placement strategy has to be deployed based on the correlation in the program tasks,the service placement strategy is formulated by considering the correlation of the program tasks.In this research,a Dynamic Service Placement List Scheduling(DSPLS)algorithm based on remaining service time prediction is proposed.The algorithm mainly optimizes the entire service placement strategy from two aspects of data transmission and service time.From the results of the simulation experiments,the transmission delay of the DSPLS algorithm is always the lowest.To complete the entire task,the service time required by DSPLS algorithm is the minimum when compared to other algorithms.2.A service placement approach for latency-sensitive applications in Mobile Edge Computing(MEC).In the MEC scenario where the user is moving,the previous optimal service placement strategy may become a non-optimal service placement strategy within a few minutes.A novel advanced genetic algorithm(Service Placement:Improved Genetic Algorithm,SPIGA)is designed for this problem.The SPIGA algorithm mainly makes service deployment decisions based on three constraints: user mobility,edge server resources,and energy costs.The goal of the SPIGA algorithm is to obtain the best Quality of Service(Qo S)under these constraints.The algorithm combines the advantages of the simulated annealing algorithm on the basis of the genetic algorithm.By introducing the idea of the simulated annealing algorithm,the genetic algorithm is easy to fall into the local optimum.The simulation results of this study show that the SPIGA algorithm can obtain higher Qo S values than other comparison algorithms.Moreover,the SPIGA algorithm makes the resource utilization of edge computing servers higher.