On-Line End-to-End Congestion Control

TL;DR

This paper analyzes an end-to-end congestion control protocol, demonstrating that it nearly maximizes total throughput under certain conditions, with explicit convergence rates and adaptability to network changes.

Contribution

It provides a rigorous analysis of a multiplicative-increase, multiplicative-decrease protocol, including convergence rates and robustness to network dynamics.

Findings

Protocol achieves near-optimal throughput when all users adopt it.

Includes explicit convergence rates for the protocol.

Handles variable round-trip-times, connection dynamics, and capacity changes.

Abstract

Congestion control in the current Internet is accomplished mainly by TCP/IP. To understand the macroscopic network behavior that results from TCP/IP and similar end-to-end protocols, one main analytic technique is to show that the the protocol maximizes some global objective function of the network traffic. Here we analyze a particular end-to-end, MIMD (multiplicative-increase, multiplicative-decrease) protocol. We show that if all users of the network use the protocol, and all connections last for at least logarithmically many rounds, then the total weighted throughput (value of all packets received) is near the maximum possible. Our analysis includes round-trip-times, and (in contrast to most previous analyses) gives explicit convergence rates, allows connections to start and stop, and allows capacities to change.