Research On Internet Content Predicting Of LTE Mobile Communication And Base Station Cache Strategy

The services provided by LTE mobile communication system(4G)have been extended to all aspects of our lives.The upcoming 5G mobile communication system(5G)will provide people with richer and better communication services with faster transmission speed,lower delay,and higher connection density.A series of application scenarios such as cities,smart homes,car networking,and internet of things will also become a reality.Realizing these scenarios will cause the backhaul connection of the key nodes of the communication network to undertake enormous data transmission pressure.If this problem cannot be solved properly,it will cause the network to have a great delay or even crash in the peak period of users.The increase in physical equipment alone cannot cope with a 100-fold increase in data traffic.The content distribution network(CDN)can implement hotspot content caching to the underlying base station of the network,allowing a large amount of highly repetitive data to be transmitted through the underlying base station,thereby alleviating the traffic pressure of the top-level backhaul connection.Based on this background,this paper studies the popularity prediction problem of content in LTE mobile communication system and the optimization of the content in the base station's cache,and proposes two content popularity prediction algorithms and two content cache algorithms.The main work of this article is as follows.1.Research on HTTP signalling data preprocessing problem.The format,properties,and characteristics of LTE signaling data are analyzed in depth,and a data preprocessing system method is provided for the characteristics of the data.The data is preprocessed from the perspectives of program design,multithreading,database construction and SQL statement optimization.The proposed method can effectively deal with and deal with large-scale,massive file abnormal data.2.Researched content popularity prediction problems.Combining with the characteristics of time series data,a content popularity prediction algorithm based on ARIMA method and a content popularity prediction algorithm based on LSTM neural network method are proposed.According to the historical request time sequence of a piece of content,the content is predicted at the next time in the whole network range.The number of requests.Also presented is a method of deriving its popularity in a single cell from the overall network popularity of the content.Numerical experiments verify the performance of the proposed algorithm.3.The problem of base station caching is studied.A 0-1 integer linear programming model of the problem is established and two solving algorithms are proposed.One is a single-base-station algorithm that only considers the content buffering of a single base station.This problem is transformed into a single-knapsack problem,and a dynamic programming algorithm is proposed.The other is a multi-base-station algorithm that considers content sharing among multiple base stations while considering the content buffering of a base station and its neighboring base stations.Through numerical experiments,the proposed algorithm and the shortage algorithm are compared and analyzed,which verifies the practicability of multi-base station algorithm.