Delay Optimal Secrecy in Two-Relay Network

TL;DR

This paper investigates delay-optimal secure communication strategies in a two-relay network with varying channel states, demonstrating that minimal feedback delay can nearly achieve the optimal secrecy delay.

Contribution

It introduces an encoding strategy for achieving optimal delay in secure relay communication under both immediate and one-block delayed feedback scenarios.

Findings

Optimal delay can be achieved with no feedback delay.

One-block delayed feedback results in at most one block delay difference.

The proposed strategies adapt to arbitrary, non-stationary channel states.

Abstract

We consider a two-relay network in which a source aims to communicate a confidential message to a destination while keeping the message secret from the relay nodes. In the first hop, the channels from the source to the relays are assumed to be block-fading and the channel states change arbitrarily -possibly non-stationary and non-ergodic- across blocks. When the relay feedback on the states of the source-to-relay channels is available on the source with no delay, we provide an encoding strategy to achieve the optimal delay. We next consider the case in which there is one-block delayed relay feedback on the states of the source-to-relay channels. We show that for a set of channel state sequences, the optimal delay with one-block delayed feedback differs from the optimal delay with no-delayed feedback at most one block.