一种基于下行控制信息的移动通信流量分类方法

A Mobile Communication Traffic Classification Method Based on Downlink Control Information

伴随着诸多新兴业务的发展,对移动通信的数据速率的要求也越来越高,网络运营商需要在确保用户隐私安全的情况下进行流量分类,分配所需的网络资源来服务用户,更好地优化移动通信的体系结构。传统的基于协议特征、关键字信息的流量分类方法会带来精度下降、实时性不足等问题。文章采集4G LTE网络中的下行控制信息(Downlink Control Information,DCI),训练了三类基准机器学习分类模型和长短期记忆网络(Long Short-Term Memory,LSTM)模型,识别了5种主流应用流量。实验结果表明,LSTM模型的F1-Score达到了98.92%,在不侵犯用户隐私的前提下,实现了移动通信流量的高精度分类。

With the growth of several developing services,the data rate needs for mobile communications are increasing.Network operators must classify traffic while maintaining user privacy and security,and then distribute the necessary network resources to provide the best possible service to consumers.The architecture of mobile communication that has been optimized.Traditional techniques of traffic classification based on protocol characteristics and keyword information will introduce issues such as decreasing accuracy and insufficient real-time performance.This paper collected Downlink Control Information(DCI)in a 4G LTE network and trained three different types of benchmark machine learning classification models and Long Short-Term Memory(LSTM)models to identify five common application traffic patterns.The results indicate that the LSTM model's F1-Score is 98.92 percent,allowing for high-precision classification of mobile communication traffic without jeopardizing user privacy.