Research On Edge Cache Deployment And Optimization Strategy In Ultra-Dense Heterogeneous Networks

With the development of communication technology,the form of data business diversification,all people have entered the Internet era,the data flow produced by intelligent devices have reached Ze bytes(ZB),huge amounts of data flow brought great pressure to the mobile network,so the equipment user request data generated when the emergence of network congestion and related requirements are inspired.An effective solution is to deploy computing and caching resources on the edge of the mobile communication network,which can provide communication and data processing services on the edge of the network closer to the users.Edge cache based on 5G communication architecture provides technical support for the development of distributed data storage.Edge caching brings remote server storage closer to the edge of the network and improves data tasks in traditional network architectures,such as distributed content caching,data prediction,and performance optimization.By storing the hot spot data information accessed by the user in advance to the local base station,the user can directly download the request from the local place and reduce the redundant data transmission in the network,so as to save the network resources and obtain a higher quality user experience.In this thesis,under the coexistence of small base stations and WiFi in 5G ultra-dense heterogeneous network,the content storage and distribution mechanism based on edge cache is designed.The main research is as follows:1.A cache strategy in a scenario where a small base station and a WiFi access point coexist in a 5G ultra-dense heterogeneous network is proposed.Considering the convenience and wi-fi access points,high bandwidth users having chosen their data request download selective connected wi-fi to reduce the traffic cost,at the same time that is likely to affect the cache performance characteristics of multiple factors(frequency of user access,file download costs,file storage space)for selecting and using AHP weight assignment algorithm and avoid the single feature factors on the deviation produced by the cache performance,caching policy for this scenario has carried on the modeling and analysis.2.According to the distributed deployment scenario of the cache file,the consideration of the file copy is introduced,and a file copy caching algorithm based on the clustering of the base station is proposed to reduce the interference between the small base stations in the ultra-dense heterogeneous network,thereby further improving the user's increase data access speed andenhance user experience.3.Aiming at the influence of geographical location and spatial storage capacity of base station deployment on base station cache performance,an algorithm based on long-short-time memory network(LSTM)for predicting base station storage file information and optimizing base station distance is proposed.The algorithm is based on the input base station.The historical data is modeled and predicted in the next step,and the spatial position and storage capacity of the base station are adjusted according to the predicted value.A deep learning algorithm model with high prediction accuracy and excellent decoupling is presented to solve the small space distance optimization and cache correlation content distribution.