Robust and Undetectable Steganographic Timing Channels for i.i.d. Traffic

摘要

Steganographic timing channels exploit inter-packet delays in network traffic to transmit secret messages. The two most important design goals are undetectability and robustness. In previous proposals undetectability has been validated only against a set of known statistical methods, leaving the resistance against possible future attacks unclear. Moreover, many existing schemes do not provide any robustness at all. In this paper, we introduce a steganographic timing channel that is both robust and provably undetectable for network traffic with independent and identically distributed (i.i.d.) inter-packet delays. I.i.d. traffic models are very useful because they are simple to analyze, and constitute essential elements of many advanced network traffic models. In contrast to previous work on i.i.d. traffic we do not rely on any strong assumptions, e.g., bounded jitter, but require only the existence of a cryptographically secure pseudorandom generator. We verify the effectiveness of our approach by conducting a series of experiments on Telnet traffic and discuss the trade off between various encoding and modulation parameters.