Research On Resource Allocation Strategies For Live Video Streaming In Mobile Edge Computing

With the development of communication technology and the increasing popularity of mobile streaming media services,live video services are becoming more and more popular with people and occupy a large portion of mobile data traffic.In a traditional cloud computing environment,in order to ensure the real-time performance of live video,a large network bandwidth needs to be allocated to users,which greatly increases the network load in the cloud environment.The emergence of mobile edge computing provides a new solution to the real-time problem of live video.By using the storage resources provided by the edge server to temporarily cache the requested video,reducing user access to the core network,it can meet more of users' live video requests.However,limited edge server resources will cause users to compete for resources,which will affect the offload of terminal applications and the resource allocation of edge servers.At the same time,user live video request routing in the overlapping area of multiple edge servers will affect the edge server's resource allocation for live video.At the same time,user mobility will also cause it to switch between different servers,which will affect the stability of live video.Aiming at the above problems,this paper aims at compromising task completion time and terminal execution energy consumption,and studies serial task offloading and resource allocation strategies with limited resources in edge computing.With the goal of maximizing the user's experience of watching live video,the research focuses Resource allocation schemes for live video streams in edge server overlay scenarios and research on resource allocation strategies for live video streams in mobile edge computing.The main results of this paper are as follows:(1)Aiming at the multi-user competition for resources caused by limited edge server resources in the edge computing environment,a multi-user serial task dynamic offloading strategy(MSTDOS)is proposed.First,the task model and completion time and energy consumption model are established by analyzing the characteristics of the serial application and the edge server.Then,an optimization goal is constructed to jointly consider the task completion time and the terminal execution energy consumption,and prove that this problem is NP-hard.Based on the chemical reaction algorithm,a heuristic algorithm is designed to solve.Simulation results show that the MSTDOS strategy can achieve better performance.2)Aiming at the problem of user request routing and resource allocation of live video by the edge server in a multi-edge server overlapping environment,a multi-edge server resource allocation scheme(MESRAS)is proposed.The server makes an approximately optimal resource allocation strategy.First,analyze the characteristics of the edge environment to establish the user request model and the storage and network resource model of the edge server;then,establish an optimization problem with the goal of maximizing the video viewing quality of all users,and convert the problem into a knapsack problem;finally,based on Greedy design user request routing algorithm and design resource allocation algorithm based on Lagrange multiplier method to solve.Simulation results show that the MESRAS scheme can achieve better user video viewing quality in a multi-edge server overlay environment.(3)Aiming at the instability of live video playback caused by users moving randomly between cells in an edge environment,a mobile-aware edge server resource allocation strategy(MAESRA)was designed.First,the user's mobility model is abstracted by analyzing the user's mobile characteristics at different time periods.Second,the storage server and network resource model of the edge server and the longterm cost model of the video provider are established.Then,the objective function is designed while taking into account the user video playback Quality and video jitter;Finally,based on Lyapunov optimization technology,an online algorithm is proposed to solve the above optimization problems.Simulation results show that the performance of MAESRA strategy is significantly better than other representative heuristic benchmark strategies.