Equilibrium Properties of Rate Control Protocols

TL;DR

This paper investigates the stability and nonlinear dynamics of the Rate Control Protocol (RCP) using models that reveal how protocol parameters and feedback forms influence network stability and lead to bifurcations.

Contribution

It provides a detailed analysis of RCP's stability, highlighting the effects of different feedback mechanisms and time scales on nonlinear and stochastic properties.

Findings

Existence of Hopf bifurcation in RCP models

Identification of stable limit cycles due to bifurcation

Impact of propagation and queuing delays on stability

Abstract

We analyze the stability of the Rate Control Protocol (RCP) using two different models that have been proposed in literature. Our objective is to better understand the impact of the protocol parameters and the effect different forms of feedback have on the stability of the network. We also highlight that different time scales, depending on the propagation delay relative to the queuing delay, have an impact on the nonlinear and the stochastic properties of the protocol fluid models. To better understand some of the nonlinear properties, we resort to local bifurcation analysis where we exhibit the existence of a Hopf type bifurcation that then leads to stable limit cycles. Our work serves as a step towards a more comprehensive understanding of the nonlinear fluid models that have been used as representative models for RCP.