A Longitudinal Analysis of Internet Rate Limitations

Year
2014
Abstract

TCP remains the dominant transport protocol for Internet traffic, but the preponderance of its congestion control mechanisms in determining flow throughput is often disputed. This paper analyzes the extent to which network, host and application settings define flow throughput over time and across autonomous systems. Drawing from a longitudinal study spanning five years of passive traces collected from a single transit link, our results show that continuing OS upgrades have reduced the influence of host limitations owing both to windowscale deployment, which by 2011 covered 80% of inbound traffic, and increased socket buffer sizes. On the other hand, we show that for this data set, approximately half of all inbound traffic remains throttled by constraints beyond network capacity, challenging the traditional model of congestion control in TCP traffic as governed primarily by loss and delay.

Summary

A five-year longitudinal study of passive traces from a single transit link examining what really limits TCP flow throughput over time: finds that OS upgrades (window-scale deployment reaching 80% of inbound traffic by 2011, larger socket buffers) have steadily reduced host-side limitations, yet roughly half of inbound traffic remains throttled by constraints beyond network capacity, challenging the traditional loss-and-delay view of TCP congestion control.

bibtex
@inproceedings{araujo2014ratelimitations,
author = {Jo\~ao Taveira Ara\'ujo and Ra\'ul Landa and Richard G. Clegg and George Pavlou and Kensuke Fukuda},
title = {A Longitudinal Analysis of Internet Rate Limitations},
booktitle = {IEEE INFOCOM 2014},
year = {2014},
pages = {1438--1446},
doi = {10.1109/infocom.2014.6848078}
}
Authors
João Taveira Araújo, Raúl Landa, Richard G. Clegg, George Pavlou, Kensuke Fukuda
Venue
IEEE INFOCOM 2014, pp. 1438-1446