A model of steganographic embedding in a message stream with error-correcting code

Abstract

Steganographic data embedding increases the security of information by hiding the fact that a message exists. All steganographic methods focus on a specific steganographic container format. Network steganography uses network packets as containers. Error-correcting codes are used to transmit messages over the network to ensure data integrity. This feature of network packets poses additional challenges for steganographic methods. This paper proposes a method to covertly embed data in a message stream with error-correcting coding. The embedded message is masked as discrete noise in the information channel. Embedding is performed by inverting individual bits in some packets. The algorithm uses pseudo-random sequence generators to select the packet number and the number of the bit to be inverted. The generator parameters are the embedding key. The receiving side retrieves the message based on the known location of the corrected bits. The embedding key also includes a start packet from which the hidden data is placed. The error-correcting code restores the original packets. Extraction of the hidden message should be performed before packet recovery. Mathematical modeling of the proposed scheme is performed and analytical regularities for the influence of embedding on message transmission over a channel with discrete noise are obtained. Probability estimation for correct transmission of packets without embedded message and with embedded data is carried out. A computer experiment for the scheme with Hamming code is performed. The experimental dependence of the throughput for the steganographic scheme and the integrity of the embedded message on the probability of error per bit in the noisy information channel is obtained. The results of computational experiment agree with analytical estimates.