Why is the internet so slow?!

Ilker Nadi Bozkurt, Anthony Aguirre, Balakrishnan Chandrasekaran, P. Brighten Godfrey, Gregory Laughlin, Bruce Maggs, Ankit Singla

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In principle, a network can transfer data at nearly the speed of light. Today’s Internet, however, is much slower: our measurements show that latencies are typically more than one, and often more than two orders of magnitude larger than the lower bound implied by the speed of light. Closing this gap would not only add value to today’s Internet applications, but might also open the door to exciting new applications. Thus, we propose a grand challenge for the networking research community: building a speed-of-light Internet. To help inform research towards this goal, we investigate, through large-scale measurements, the causes of latency inflation in the Internet across the network stack. Our analysis reveals an under-explored problem: the Internet’s infrastructural inefficiencies. We find that while protocol overheads, which have dominated the community’s attention, are indeed important, reducing latency inflation at the lowest layers will be critical for building a speed-of-light Internet. In fact, eliminating this infrastructural latency inflation, without any other changes in the protocol stack, could speed up small object fetches by more than a factor of three.

Original languageEnglish (US)
Title of host publicationPassive and Active Measurement - 18th International Conference, PAM 2017, Proceedings
PublisherSpringer Verlag
Pages173-187
Number of pages15
Volume10176 LNCS
ISBN (Print)9783319543277
DOIs
StatePublished - 2017
Event18th International Conference on Passive and Active Measurement, PAM 2017 - Sydney, Australia

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10176 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other18th International Conference on Passive and Active Measurement, PAM 2017
CountryAustralia
City Sydney
Period3/30/173/31/17

Fingerprint

World Wide Web
Internet
Latency
Light velocity
Inflation
Data transfer
Networking
Lowest
Speedup
Lower bound

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Bozkurt, I. N., Aguirre, A., Chandrasekaran, B., Godfrey, P. B., Laughlin, G., Maggs, B., & Singla, A. (2017). Why is the internet so slow?! In Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings (Vol. 10176 LNCS, pp. 173-187). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10176 LNCS). Springer Verlag. DOI: 10.1007/978-3-319-54328-4_13

Why is the internet so slow?! / Bozkurt, Ilker Nadi; Aguirre, Anthony; Chandrasekaran, Balakrishnan; Godfrey, P. Brighten; Laughlin, Gregory; Maggs, Bruce; Singla, Ankit.

Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings. Vol. 10176 LNCS Springer Verlag, 2017. p. 173-187 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10176 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Bozkurt, IN, Aguirre, A, Chandrasekaran, B, Godfrey, PB, Laughlin, G, Maggs, B & Singla, A 2017, Why is the internet so slow?! in Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings. vol. 10176 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10176 LNCS, Springer Verlag, pp. 173-187, 18th International Conference on Passive and Active Measurement, PAM 2017, Sydney, Australia, 30-31 March. DOI: 10.1007/978-3-319-54328-4_13
Bozkurt IN, Aguirre A, Chandrasekaran B, Godfrey PB, Laughlin G, Maggs B et al. Why is the internet so slow?! In Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings. Vol. 10176 LNCS. Springer Verlag. 2017. p. 173-187. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Available from, DOI: 10.1007/978-3-319-54328-4_13

Bozkurt, Ilker Nadi; Aguirre, Anthony; Chandrasekaran, Balakrishnan; Godfrey, P. Brighten; Laughlin, Gregory; Maggs, Bruce; Singla, Ankit / Why is the internet so slow?!

Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings. Vol. 10176 LNCS Springer Verlag, 2017. p. 173-187 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10176 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

@inbook{b1fa7d896b0c4ea4a0010028763edf5c,
title = "Why is the internet so slow?!",
abstract = "In principle, a network can transfer data at nearly the speed of light. Today’s Internet, however, is much slower: our measurements show that latencies are typically more than one, and often more than two orders of magnitude larger than the lower bound implied by the speed of light. Closing this gap would not only add value to today’s Internet applications, but might also open the door to exciting new applications. Thus, we propose a grand challenge for the networking research community: building a speed-of-light Internet. To help inform research towards this goal, we investigate, through large-scale measurements, the causes of latency inflation in the Internet across the network stack. Our analysis reveals an under-explored problem: the Internet’s infrastructural inefficiencies. We find that while protocol overheads, which have dominated the community’s attention, are indeed important, reducing latency inflation at the lowest layers will be critical for building a speed-of-light Internet. In fact, eliminating this infrastructural latency inflation, without any other changes in the protocol stack, could speed up small object fetches by more than a factor of three.",
author = "Bozkurt, {Ilker Nadi} and Anthony Aguirre and Balakrishnan Chandrasekaran and Godfrey, {P. Brighten} and Gregory Laughlin and Bruce Maggs and Ankit Singla",
year = "2017",
doi = "10.1007/978-3-319-54328-4_13",
isbn = "9783319543277",
volume = "10176 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "173--187",
booktitle = "Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings",
address = "Germany",

}

TY - CHAP

T1 - Why is the internet so slow?!

AU - Bozkurt,Ilker Nadi

AU - Aguirre,Anthony

AU - Chandrasekaran,Balakrishnan

AU - Godfrey,P. Brighten

AU - Laughlin,Gregory

AU - Maggs,Bruce

AU - Singla,Ankit

PY - 2017

Y1 - 2017

N2 - In principle, a network can transfer data at nearly the speed of light. Today’s Internet, however, is much slower: our measurements show that latencies are typically more than one, and often more than two orders of magnitude larger than the lower bound implied by the speed of light. Closing this gap would not only add value to today’s Internet applications, but might also open the door to exciting new applications. Thus, we propose a grand challenge for the networking research community: building a speed-of-light Internet. To help inform research towards this goal, we investigate, through large-scale measurements, the causes of latency inflation in the Internet across the network stack. Our analysis reveals an under-explored problem: the Internet’s infrastructural inefficiencies. We find that while protocol overheads, which have dominated the community’s attention, are indeed important, reducing latency inflation at the lowest layers will be critical for building a speed-of-light Internet. In fact, eliminating this infrastructural latency inflation, without any other changes in the protocol stack, could speed up small object fetches by more than a factor of three.

AB - In principle, a network can transfer data at nearly the speed of light. Today’s Internet, however, is much slower: our measurements show that latencies are typically more than one, and often more than two orders of magnitude larger than the lower bound implied by the speed of light. Closing this gap would not only add value to today’s Internet applications, but might also open the door to exciting new applications. Thus, we propose a grand challenge for the networking research community: building a speed-of-light Internet. To help inform research towards this goal, we investigate, through large-scale measurements, the causes of latency inflation in the Internet across the network stack. Our analysis reveals an under-explored problem: the Internet’s infrastructural inefficiencies. We find that while protocol overheads, which have dominated the community’s attention, are indeed important, reducing latency inflation at the lowest layers will be critical for building a speed-of-light Internet. In fact, eliminating this infrastructural latency inflation, without any other changes in the protocol stack, could speed up small object fetches by more than a factor of three.

UR - http://www.scopus.com/inward/record.url?scp=85015895517&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015895517&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-54328-4_13

DO - 10.1007/978-3-319-54328-4_13

M3 - Conference contribution

SN - 9783319543277

VL - 10176 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 173

EP - 187

BT - Passive and Active Measurement - 18th International Conference, PAM 2017, Proceedings

PB - Springer Verlag

ER -