基于深度学习的加密网络流量识别方法研究  

Encrypted Network Traffic Identification Method based on Deep Learning

作　　者：吉庆兵[1,2] 陈江涛 潘炜 JI Qingbing;CHEN Jiangtao;PAN Wei(No.30 Institute,CETC,Chengdu Sichuan 610041,China;School of Cybersecurity,Northwestern Polytechnical University,Xi’an Shaanxi 710072,China;School of Computer Science,Northwestern Polytechnical University,Xi’an Shaanxi 710072,China)

机构地区：[1]中国电子科技集团公司第三十研究所,四川成都610041 [2]西北工业大学网络空间安全学院,陕西西安710072 [3]西北工业大学计算机学院,陕西西安710072

出　　处：《通信技术》2021年第1期152-157,共6页Communications Technology

基　　金：保密通信重点实验室基金项目(No.6142103190308)。

摘　　要：随着网络技术的快速发展以及TLS、SSL、SSH等加密技术在网络中的广泛使用,网络加密流量快速增长,加密俨然已经成为保护隐私的重要手段之一。但是,经过加密的流量在无意间也给网络安全带来了新的隐患。加密网络流量的分析识别可以有效防范恶意流量,对保证网络信息安全和维护网络正常运行具有重要实际意义。提出了一种基于深度学习的加密网络流量识别方法,并进行了实验验证,结果表明,该方法在14类网络应用的加密流量上的识别准确率为96.22%,能够满足实际需求。With the rapid development of network technology and the widespread use of encryption technologies such as TLS,SSL,SSH in the network,network encryption traffic has grown rapidly,and encryption has become one of the important means to protect privacy.However,the encrypted traffic unintentionally also brings new hidden dangers to network security.The analysis and identification of encrypted network traffic can effectively prevent malicious traffic,which has important practical significance for ensuring network information security and maintaining the normal operation of the network.An encrypted-network-traffic identification mebthod based on deep learning is proposed,and experimental verification indicates that the encrypted-traffic identification accuracy of this method on 14 types of network application is 96.22%,which can meet actual needs.

关 键 词：网络安全 网络流量识别 深度学习 深度自编码器 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]