Research And Prototype Verification Of Edge Network Caching Technology Based On SDN

With the rise of new network applications,such as Internet of Vehicles,Smart City,Virtual Reality and so on,the type of service in the edge wireless network is increasing.Differentiated services have different requirements for delay,which brings new challenges to edge caching technology.In this thesis,we propose a low latency edge wireless network caching system based on Software Defined Network(SDN),which aims to reduce the user's visiting delay and downloading delay.Visiting delay is the time it takes for a user to visiting the node storing the required content.The delay is mainly determined by the location of the content.Downloading delay indicates the amount of time the content is sent back to the user from the cache node,which is mainly determined by the maximum transmission rate(i.e.throughput)provided by the base station associated with the user.Therefore,this thesis mainly focuses on user association and cache decision.(1)Association decision model based on load balance.In the traditional RSSI-based association decision,the user selects the base station from the perspective of its own,does not consider the load of the base station and the overall performance of the network,which will undoubtedly reduce the network system throughput,increase the average user downloading delay?Therefore,the association decision not only need to ensure the relative balance of the base station load,but also need to maximize system throughput.Based on this goal,this thesis proposes an association decision model based on load balance.However,with the expansion of network size,the association decision model is much difficult to solve.In order to reduce the difficulty of solving,this thesis proposes a KLA-based association algorithm.The algorithm creates virtual base station by introducing graph partitioning.Therefore,the user first needs to access the virtual base station and then associate with a specific physical base station in each virtual base station.(2)Cache decision model based on network slice.There are usually multiple service providers in an edge wireless network,and different service providers have different requirements for access latency.To promote the user experience,cache decision need to reduce the average visiting delay in the network and meet the visiting delay requirements of different service providers.Based on this goal,this thesis virtualizes the cache resource to meet the vistiting delay requirements of different services by introducing the network slice,and proposes a cache decision model based on network slicing.At the same time,this thesis proves that the cache decision model has NP completeness,and design a heuristic algorithm to solve the model.Based on the above research,a lot of simulations have been carried out to verify the performance of the association decision model based on load balance and cache decision model based on network slicing.Finally,this thesis uses the onboard wireless server to prototype the entire cache system.