Research On The Wireless Adaptive Video Streaming Transmission Technology Based On Edge Computing

With the rapid development of mobile Internet and the popularity of video services,the traffic generated by video services has occupied most of the current network traffic.To enhance the user experience,a wide variety of mobile terminals also support video viewing.However,the diversity of terminals,the diversity of user requirements and the time-varying channel bring great challenges to video transmission.At present,Cloud-based Dash System can deal with these problems well,which has been widely deployed to the actual system.However,the system still has the following four shortcomings: First,in traditional DASH system,pre-fixed bitrate set is used to deal with the time-varying network conditions and users' demands in all kinds of scenarios,which is inflexible.Second,each version of the same video has to be transmitted repeatedly between the core network and the base station,which will waste the bandwidth resources.Third,servers in remote clouds can cause unacceptable delays when serving users.Fourth,the core network server is responsible for the transcoding and storage services of all cell users' video,which makes the computing and storage pressure on the server too great.Because most of the users using video services are mobile users and the channel changes of mobile users are faster and more complex,it is necessary to design a new system architecture or algorithm to solve the above problems.In order to solve these problems,this thesis introduces the concept of edge computing,and proposes an adaptive video transmission system based on edge computing,which sets the transcoding server on the edge of the base station.In our proposed system,the core network only needs to transmit the source video stream once to the edge transcoding server,and the transcoding server will transcode the best bitrate set according to the needs of the users and the network conditions,which improves the disadvantages of the traditional DASH system.In addition,in order to improve bandwidth utilization and further improve system performance,this thesis also implements the wireless spectrum(bandwidth)adjustment function on the edge server.We use the idea of cross-layer optimization(application-physical layer)to model the adaptive video transmission policy problem in the new architecture as a network utility maximization problem and furtherdecompose it into two sub-problems,namely,the bitrate allocation problem and the bandwidth adjustment problem.The analyses prove that the bitrate allocation problem is the NP-Hard problem,and there is no approximation algorithm in polynomial time.Therefore,this thesis proposes two heuristic algorithms to solve it.For the bandwidth adjustment problem,we prove which is a DC(Difference of Convex)programming problem.We propose a low complexity iterative algorithm based on CCCP(Concave-Convex Procedure)to solve the bandwidth adjustment problem.In order to make the algorithm enter the iterative calculation with a feasible initial point,we also propose an iterative feasible initial point search algorithm to find a feasible point for the original DC programming problem.These two algorithms can be combined into a "two step algorithm" to solve the original DC problem.A large number of simulation experiments demonstrate the advantages of the proposed system compared with the traditional DASH system,as well as the feasibility and computational complexity of the rate allocation algorithm and bandwidth adjustment algorithm designed by us.The simulation results show that the performance of the system is far superior to the traditional system,and the computational complexity and the convergence of the bandwidth adjustment algorithm are all acceptable.