Secure Error-Correction Network Coding in a Randomized Setting

Yejun Zhou, Hui Li, Jianfeng Ma

Abstract: Using a metric on the space of all subspaces of a fixed vector space, a secure error-correcting network coding scheme is proposed. Firstly, We show that, a limited dimension errors can be corrected using the minimum distance decoder for this metric. Secondly, when the number of channels the adversary can eavesdrop on is limited (two types of adversaries is considered), he can’t get any information about the source. Finally, when the source and the destination can share a low rate secret channel and a Pseudo-Random number generator, we present a coding scheme which can prevent eavesdropping adversaries who can eavesdrop on all the channels of a network.