计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (2): 68-75.DOI: 10.3778/j.issn.1002-8331.1811-0076

• 网络、通信与安全 • 上一篇    下一篇

差分隐私保护的Android应用流量行为混淆方法

王佳贺,魏松杰,吴超   

  1. 南京理工大学 计算机科学与工程学院,南京 210094
  • 出版日期:2020-01-15 发布日期:2020-01-14

Android Apps Traffic Behavior Obfuscation Method Based on Differential Privacy Protection

WANG Jiahe, WEI Songjie, WU Chao   

  1. School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing 210094, China
  • Online:2020-01-15 Published:2020-01-14

摘要: 针对Android用户在终端传输数据和发送信息所带来的网络应用行为暴露等问题,通过结合自定义流量混淆和差分隐私无关流量干扰两种方法的优势,能够在保证Android应用网络连接状态和数据传输内容不变的前提下,通过改变流量数据包的时序和数目特征,达到对指定用户应用行为特征的隐私保护。实验结果表明,选取Android典型应用流量并提取六种主要流量特征,对比混淆前后数据包特征,所提混淆方法能够有效地改变Android终端的应用流量,抵御支持向量机(Support Vector Machine,SVM)和BP(Back Propagation)神经网络算法的识别,准确率高达96.55%,最终实现对Android终端应用行为的保护。

关键词: 安卓, 差分隐私, 流量混淆, 隐私保护

Abstract: To overcome the difficulty in network application behavior exposure caused by Android users transmitting data and sending information through the terminal, this paper combines the advantages of two methods, namely, custom traffic confusion method and differential privacy-independent traffic interference, which can realize the implement privacy protection for the characteristics of user application behavior by changing the timing sequence and number features of traffic packets while ensuring the network connection status and data transmission content of the Android applications. The experimental results show that by selecting typical Android application traffic and extracting six main kinds of traffic characteristics, comparing the packet characteristics before and after confusion, the confusion method of this paper can effectively change the application traffic of Android terminals and resist the recognition of Support Vector Machine(SVM) and BP(Back Propagation) neural network algorithms, whose accuracy rate is as high as 96.55%, and finally realizes the protection of application behavior of Android terminals.

Key words: Android, differential privacy, traffic obfuscation, privacy protection