An Analytical Model for the Partial Intercept Probability in Sparse Linear Network Coding

Abstract

The security of linear network coding schemes such as Random Linear Network Coding (RLNC) and Sparse Random Linear Network Coding (SRLNC) is an important performance metric. One of the security aspects of these coding schemes is the probability that a potential eavesdropper recovers a fraction of source packets. In this work, we consider a network consisted of a sender, a legitimate receiver and an eavesdropper, where the sender uses SRLNC to broadcast data. We propose an analytical approximation for the probability of decoding a fraction of source packets, i.e., the partial intercept probability, by the eavesdropper. Using this analytical model, we propose an algorithm for the maximum sparsity that satisfies a threshold on the number of the source packets decoded by the eavesdropper. Using simulation technique, We proved that the maximum sparsity found by this algorithm satisfies the aforementioned threshold.