Multiple-Relay Slotted ALOHA: Performance Analysis and Bounds

TL;DR

This paper analyzes the performance of multi-receiver slotted ALOHA protocols in wireless communications, deriving exact throughput and loss rate expressions, and establishing bounds on downlink resources with practical forwarding policies.

Contribution

It provides exact analytical expressions for throughput and packet loss in multi-receiver ALOHA, and introduces bounds and policies for efficient resource utilization.

Findings

Exact uplink throughput and loss rate formulas for any number of relays.

A lower bound on downlink resources needed for complete data delivery.

Performance characterization of simple forwarding policies with finite buffers.

Abstract

Wireless random access protocols are attracting a revived research interest as a simple yet effective solution for machine-type communications. In the quest to improve reliability and spectral efficiency of such schemes, the use of multiple receivers has recently emerged as a promising option. We study the potential of this approach considering a population of users that transmit data packets following a simple slotted ALOHA policy to a set of non-cooperative receivers or relays (uplink phase). These, in turn, independently forward - part of - what decoded towards a collecting sink (downlink phase). For an on-off fading channel model, we provide exact expressions for uplink throughput and packet loss rate for an arbitrary number of relays, characterising the benefits of multi-receiver schemes. Moreover, a lower bound on the minimum amount of downlink resources needed to deliver all information collected on the uplink is provided. The bound is proven to be achievable via random linear coding when no constraints in terms of latency are set. We complement our study discussing a family of simple forwarding policies that require no packet-level coding, and optimising their performance based on the amount of available downlink resources. The behaviour of both random linear coding and simplified policies is also characterised when receivers are equipped with finite buffers, revealing non-trivial tradeoffs.