Instability of Sharing Systems in the Presence of Retransmissions

TL;DR

This paper reveals that processor sharing scheduling becomes completely unstable with zero throughput in failure-prone networks due to retransmissions, advocating for simpler scheduling policies like first-come-first-serve for stability.

Contribution

It uncovers a new instability phenomenon in sharing systems caused by retransmissions, showing that PS-based scheduling fails regardless of low traffic loads.

Findings

PS scheduling induces instability with zero throughput in retransmission scenarios.

Scheduling one job at a time ensures stability in failure-prone networks.

Simulation results validate the analytical instability findings.

Abstract

Retransmissions represent a primary failure recovery mechanism on all layers of communication network architecture. Similarly, fair sharing, e.g. processor sharing (PS), is a widely accepted approach to resource allocation among multiple users. Recent work has shown that retransmissions in failure-prone, e.g. wireless ad hoc, networks can cause heavy tails and long delays. In this paper, we discover a new phenomenon showing that PS-based scheduling induces complete instability with zero throughput in the presence of retransmissions, regardless of how low the traffic load may be. This phenomenon occurs even when the job sizes are bounded/fragmented, e.g. deterministic. Our analytical results are further validated via simulation experiments. Moreover, our work demonstrates that scheduling one job at a time, such as first-come-first-serve, achieves stability and should be preferred in these systems.