Cellular Network Traffic Monitoring And Cache Service Optimazation Based On Mobile Edge Computing

The rapid growth of Internet services and the popularity of smart mobile devices have led to exponential growth of global mobile traffic,while future cellular networks should provide network with high reliability,high bandwidth and low latency.In response to this demand,Mobile Edge Computing(MEC)was proposed and developed into a key technology for future mobile networks.Local processing of the traffic of the edge network through the MEC will greatly reduce the backhaul pressure of the core network and reduce the service delay.At the same time,the MEC platform will open the edge network information to the service developer,which can be used for real-time network context-related Business optimization.The main application scenarios of MEC include high-definition video on demand,AR/VR,car networking,etc.,and the sinking of content delivery network(CDN)through MEC is also of great value,so how to use MEC platform for mobile service optimization has become a research hotspot.Based on the MEC architecture,this paper studies how to monitor traffic on the edge network and open out network information for service optimization.On the other hand,the MEC is combined with the CDN service to realize the sinking of the CDN node to the edge of the mobile network,to solve the problem of the failure of the traditional CDN to accelerate the mobile service,and further study the collaborative caching strategy of the edge CDN.The main work of the thesis includes:1.Research and implementation of edge traffic monitoring system based on MEC platform.Network traffic monitoring and opening is one of the core functions of MEC platform.This paper studies how to provide network traffic monitoring capability for MEC platform for MEC specification and system architecture released by ETSI.Open the context information on the wireless network side and provide a business-optimized RESTFUL API interface for third-party services.This paper proposes a cellular traffic monitoring method based on deep packet inspection.It can detect GTP and S1AP packets on the S1 interface in LTE network in real time,analyze the traffic and process the encapsulation into an interface to open out network information.This paper elaborates on the flow collection,data processing and storage architecture design of the traffic monitoring framework,and performs unit tests on the system framework.2.The research and implementation of the mobile edge CDN system.Traditional CDN is deployed outside the mobile network,and the acceleration to mobile services is very limited.To solve this problem,this paper explores the implementation of edge CDN acceleration service based on MEC platform,and sinks the CDN node from the traditional computer network to the edge of the cellular network.Firstly,based on the MEC platform,the edge CDN system architecture is designed.The BIND-based domain name resolution service is built.The Squid-based reverse proxy service is built to provide cache acceleration support.Finally,the performance of the edge CDN and the traditional CDN were tested in maj or application scenarios such as large file transfer,small file transfer,and audio and video on demand.Compared with traditional CDN,edge CDN can reduce DNS time and connection establishment time by 36%and 33%,respectively.3.Design and implementation of edge CDN collaborative caching system based on multi-level heterogeneous network:This paper proposes and implements a collaborative edge CDN acceleration system.Firstly,a collaborative cache architecture of adjacent nodes is proposed.Each cache node performs cache update through information sharing,which reduces cache redundancy and improves resource utilization in collaborative collection content.Then,a collaborative cache update strategy based on G-S(Gale-Shapley)algorithm is proposed,which satisfies the content and cache node preferences to optimize the overall service delay.Finally,based on the SSM framework,a collaborative cache acceleration system is built.The functions of the proposed system include collaborative node management,cache management,collaborative cache update,and high-definition video acceleration comparison.The collaborative mechanism and performance advantages of the cache system are further demonstrated through the visualized web interface.After verification,the service latency of the G-S collaborative caching strategy is reduced by 40.68%compared to the non-cooperative caching strategy,and the average popularity collaboration caching strategy is reduced by 10.01%.