Queue-Architecture and Stability Analysis in Cooperative Relay Networks

TL;DR

This paper introduces a new queue-architecture framework for cooperative relay networks that integrates information-theoretic coding with network-stability analysis, accommodating various cooperative strategies and proving throughput optimality.

Contribution

It develops a low-complexity queue-architecture and proves the throughput optimality of a network algorithm for cooperative relay networks under realistic conditions.

Findings

Proposes a novel queue-architecture for cooperative relay networks.

Establishes throughput optimality of a simple network algorithm.

Supports arbitrary cooperative coding strategies.

Abstract

An abstraction of the physical layer coding using bit pipes that are coupled through data-rates is insufficient to capture notions such as node cooperation in cooperative relay networks. Consequently, network-stability analyses based on such abstractions are valid for non-cooperative schemes alone and meaningless for cooperative schemes. Motivated from this, this paper develops a framework that brings the information-theoretic coding scheme together with network-stability analysis. This framework does not constrain the system to any particular achievable scheme, i.e., the relays can use any cooperative coding strategy of its choice, be it amplify/compress/quantize or any alter-and-forward scheme. The paper focuses on the scenario when coherence duration is of the same order of the packet/codeword duration, the channel distribution is unknown and the fading state is only known causally. The main contributions of this paper are two-fold: first, it develops a low-complexity queue-architecture to enable stable operation of cooperative relay networks, and, second, it establishes the throughput optimality of a simple network algorithm that utilizes this queue-architecture.