摘要
为识别链路层比特流是否加密,以未加密与加密数据在随机统计特性上的差异为依据,利用随机性检测的方法对加密比特流进行识别。在不同比特流长度条件下对四种典型随机性检测方法的识别率进行了比较研究。针对用户数据对未加密比特流识别率影响较大的问题,提出了基于分块游程检测及块内码元频数方差的用户数据起始位估计方案,并对提出的方案进行了评估。实验结果表明,提出的方案可显著将剔除用户数据后的未加密比特流识别率由41.23%提高到92%以上。
To identify the encrypted bit stream of data link layer, this paper used randomness tests based on the difference of statistical characteristics of unencrypted and encrypted bit stream. It compared four typical randomness tests under different length of bit stream. As the user data affected the identification rate, it proposed and assessed estimation schemes of user data＇s start bit based on run test within blocks and variance of bit frequency within blocks. Experimental results show that the proposed scheme can significantly improve the identification rate of unencrypted bit stream after removing the user data from 41.23 percent to over 92 percent.
出处
《计算机应用研究》
CSCD
北大核心
2015年第11期3443-3447,共5页
Application Research of Computers
基金
军内科研资助项目(YJJXM12033)
关键词
链路层
加密比特流
随机性检测
用户数据
data link layer
encrypted bit stream
randomness test
user data
