Vulnerability of MRD-Code-based Universal Secure Network Coding against Stronger Eavesdroppers

TL;DR

This paper demonstrates that the universal secure network coding scheme based on MRD codes by Silva et al. cannot guarantee security against adaptive eavesdroppers who can re-select tapped links during transmission, highlighting limitations in its robustness.

Contribution

The paper proves the impossibility of universal security for Silva et al.'s scheme against stronger, adaptive eavesdroppers and analyzes conditions under restricted tapping scenarios.

Findings

Universal security cannot be achieved against adaptive eavesdroppers.

Re-selection of tapped links compromises security.

Conditions for security under restricted tapping are identified.

Abstract

Silva et al. proposed a universal secure network coding scheme based on MRD codes, which can be applied to any underlying network code. This paper considers a stronger eavesdropping model where the eavesdroppers possess the ability to re-select the tapping links during the transmission. We give a proof for the impossibility of attaining universal security against such adversaries using Silva et al.'s code for all choices of code parameters, even with restricted number of tapped links. We also consider the cases with restricted tapping duration and derive some conditions for this code to be secure.