Research On Streaming And Caching Techonlogy For Mobile Edge Technology

The explosive growth of data traffic and the intensive deployment of base stations in mobile communication networks bring great pressure to the backhaul link.Mobile edge computing(MEC)alleviates the network pressure by providing computing,caching and other services at the network edge.It is considered to be one of the key technologies of the next generation mobile network.At present,the research on MEC technology focuses on theoretical analysis and simulation verification,while the research on building an experimental platform for MEC technology verification is less.At the same time,based on the MEC experimental platform,experimental verification research is more meaningful for the actual deployment of MEC in mobile networks in the future.Therefore,this paper studies MEC technology based on OAI(Open Air Interface),the open source platform of mobile communication network.The main contents include the design of shunting algorithm,the design and performance analysis of cache algorithm,and the construction and function test of MEC experimental system.In the design of shunting algorithm,S1 interface shunting algorithm based on IP quintuple rule matching(S1-IPQM)is proposed.Firstly,based on the characteristics of S1 interface protocol stack,a frame format conversion method suitable for different devices in MEC system is proposed.Secondly,the diversion rules are formulated according to the type of user business data.Finally,the S1-IPQM shunting algorithm is developed using the data plane development kit(DPDK).In the design and performance analysis of caching algorithm,considering the user’s preference and the capacity of caching system,this paper proposes a caching algorithm based on video file popularity and size(PAS).PAS caching algorithm uses video popularity and video file size to build an objective function,and updates the video in the caching system according to the objective function.Aiming at the acquisition method of video popularity in PAS cache algorithm,a Zipf parameter estimation algorithm based on maximum likelihood method is proposed.Aiming at the method of obtaining the size of video file in PAS cache algorithm,a method based on HTTP packet parsing is proposed.Finally,the PAS caching algorithm is simulated and compared with the caching algorithm based on video file popularity,the caching algorithm based on video file size and the random caching algorithm.The experimental results show that the cache hit rate of the proposed cache algorithm is higher than that of the other three cache algorithms in the case of different cache system capacity and different Zipf distribution.In terms of construction and function test of MEC experimental system,MEC experimental system is built based on OAI platform.MEC system includes base station,core network,domain name system(DNS),video server,MEC controller and user equipment(UE).The S1-IPQM shunting algorithm and PAS caching algorithm are deployed to the MEC experimental system to test the function of the system.The test results show that the MEC experimental system has the functions of shunting and caching.