Author Login Editor-in-Chief Peer Review Editor Work Office Work

Computer Engineering ›› 2019, Vol. 45 ›› Issue (12): 119-126. doi: 10.19678/j.issn.1000-3428.0053315

Previous Articles     Next Articles

Network Traffic Feature Camouflage Technology Based on Generating Adversarial Network

LI Jie, ZHOU Lu, LI Huaxin, YAN Lu, ZHU Haojin   

  1. School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
  • Received:2018-12-06 Revised:2019-01-11 Published:2019-01-04

基于生成对抗网络的网络流量特征伪装技术

李杰, 周路, 李华欣, 闫璐, 朱浩瑾   

  1. 上海交通大学 电子信息与电气工程学院, 上海 200240
  • 作者简介:李杰(1995-),男,硕士研究生,主研方向为网络安全、隐私保护、匿名网络;周路,博士研究生;李华欣,硕士研究生;闫璐,本科生;朱浩瑾(通信作者),教授、博士。
  • 基金资助:
    国家自然科学基金面上项目(61672350);装备预研教育部联合基金。

Abstract: Internet service providers often use traffic analysis attack technology to classify and monitor traffic.However,most of the current defense measures lack the flexibility to respond to changes in the network environment,and their dependence on specific traffic characteristics makes them easy to be detected and shielded.To defend against these attacks,we propose a new traffic hiding technology,which enables the proxy server dynamically camouflages any target traffic.We use the Generating Adversarial Network(GAN) model to learn the characteristics of the target traffic,making the source traffic indistinguishable from the target normal traffic so as to protect privacy and avoid Internet monitoring.Experimental results show that the proposed technology has better accuracy and flexibility,and the anonymity of traffic is obviously improved under the existing traffic analysis attack technology.

Key words: Generating Adversarial Network(GAN), network security, Website fingerprint, traffic analysis, traffic feature camouflage

摘要: 互联网服务提供商通过使用流量分析攻击技术对流量进行分类和监控,目前的防御措施多数缺乏动态性与应对网络环境变化的灵活性,而且依赖于具体的某种流量的特征,容易被探测和屏蔽。为抵御流量分析攻击,提出一种新的流量隐藏技术,代理服务器将流量动态地伪装成任意目标流量,运用生成对抗网络模型来学习目标网络流量的特征,将源流量变为与目标正常流量不可区分的流量,以保障隐私和规避互联网监控。实验结果表明,该技术在准确率等性能指标上表现良好,且拥有同类方案所不具备的动态性,明显提升了流量的匿名性。

关键词: 生成对抗网络, 网络安全, 网站指纹, 流量分析, 流量特征伪装

CLC Number: