Bandwidth sharing networks with priority scaling

TL;DR

This paper analyzes bandwidth sharing in multi-class networks with priority scaling, characterizing the dynamic interactions during traffic surges and proposing a differential equation model to improve quality of service.

Contribution

It introduces a differential equation framework for modeling bandwidth sharing with priority scaling during traffic surges, capturing the interaction between stable and unstable classes.

Findings

Differential equations describe the evolution of unstable classes during surges.

Stability regions are characterized for monotone networks.

A penalization rule is proposed for networks with streaming and elastic traffic.

Abstract

In multi-class communication networks, traffic surges due to one class of users can significantly degrade the performance for other classes. During these transient periods, it is thus of crucial importance to implement priority mechanisms that conserve the quality of service experienced by the affected classes, while ensuring that the temporarily unstable class is not entirely neglected. In this paper, we examine the complex interaction occurring between several classes of traffic when classes obtain bandwidth proportionally to their incoming traffic. We characterize the evolution of the network from the moment the initial surge takes place until the system reaches its equilibrium. Using an appropriate scaling, we show that the trajectories of the temporarily unstable class can be described by a differential equation, while those of the stable classes retain their stochastic nature. A stochastic averaging phenomenon occurs and the dynamics of the temporarily unstable and the stable classes continue to influence one another. We further proceed to characterize the obtained differential equations and the stability region under this scaling for monotone networks. We illustrate these result on several toy examples and we finally build a penalization rule using these results for a network integrating streaming and elastic traffic.