Reducing Internet Latency: A Survey of Techniques and their Merit

Bob Briscoe, Anna Brunstrom, Andreas Petlund, David Hayes, David Ross, Ing-Jyh Tsan, Stein Gjessing, Gorry Fairhurst, Carsten Griwodz, Michael Welzl

Research output: Contribution to journalArticle

53 Citations (Scopus)
354 Downloads (Pure)

Abstract

Latency is increasingly becoming a performance bottleneck for Internet Protocol (IP) networks, but historically networks have been designed with aims of maximizing throughput and utilization. This article offers a broad survey of techniques aimed at tackling latency in the literature up to March 2014 and their merits. A goal of this work is to be able to quantify and compare the merits of the different Internet latency reducing techniques, contrasting their gains in delay reduction versus the pain required to implement and deploy them. We found that classifying techniques according to the sources of delay they alleviate provided the best insight into the following issues: 1) the structural arrangement of a network, such as placement of servers and suboptimal routes, can contribute significantly to latency; 2) each interaction between communicating endpoints adds a Round Trip Time (RTT) to latency, especially significant for short flows; 3) in addition to base propagation delay, several sources of delay accumulate along transmission paths, today intermittently dominated by queuing delays; 4) it takes time to sense and use available capacity, with overuse inflicting latency on other flows sharing the capacity; and 5) within end systems delay sources include operating system buffering, head-of-line blocking, and hardware interaction. No single source of delay dominates in all cases, and many of these sources are spasmodic and highly variable. Solutions addressing these sources often both reduce the overall latency and make it more predictable.
Original languageEnglish
Pages (from-to)2149 - 2196
Number of pages49
Journal IEEE Communications Surveys & Tutorials
Volume18
Issue number3
Early online date26 Nov 2014
DOIs
Publication statusE-pub ahead of print - 26 Nov 2014

Fingerprint

Internet protocols
Servers
Throughput
Internet
Hardware

Cite this

Briscoe, B., Brunstrom, A., Petlund, A., Hayes, D., Ross, D., Tsan, I-J., ... Welzl, M. (2014). Reducing Internet Latency: A Survey of Techniques and their Merit. IEEE Communications Surveys & Tutorials, 18(3), 2149 - 2196. https://doi.org/10.1109/COMST.2014.2375213

Reducing Internet Latency : A Survey of Techniques and their Merit. / Briscoe, Bob; Brunstrom, Anna; Petlund, Andreas; Hayes, David; Ross, David; Tsan, Ing-Jyh ; Gjessing, Stein ; Fairhurst, Gorry; Griwodz, Carsten; Welzl, Michael.

In: IEEE Communications Surveys & Tutorials, Vol. 18, No. 3, 26.11.2014, p. 2149 - 2196.

Research output: Contribution to journalArticle

Briscoe, B, Brunstrom, A, Petlund, A, Hayes, D, Ross, D, Tsan, I-J, Gjessing, S, Fairhurst, G, Griwodz, C & Welzl, M 2014, 'Reducing Internet Latency: A Survey of Techniques and their Merit' IEEE Communications Surveys & Tutorials, vol. 18, no. 3, pp. 2149 - 2196. https://doi.org/10.1109/COMST.2014.2375213
Briscoe B, Brunstrom A, Petlund A, Hayes D, Ross D, Tsan I-J et al. Reducing Internet Latency: A Survey of Techniques and their Merit. IEEE Communications Surveys & Tutorials. 2014 Nov 26;18(3):2149 - 2196. https://doi.org/10.1109/COMST.2014.2375213
Briscoe, Bob ; Brunstrom, Anna ; Petlund, Andreas ; Hayes, David ; Ross, David ; Tsan, Ing-Jyh ; Gjessing, Stein ; Fairhurst, Gorry ; Griwodz, Carsten ; Welzl, Michael. / Reducing Internet Latency : A Survey of Techniques and their Merit. In: IEEE Communications Surveys & Tutorials. 2014 ; Vol. 18, No. 3. pp. 2149 - 2196.
@article{9d0bf5a159904fc9bb25db4959e11ab5,
title = "Reducing Internet Latency: A Survey of Techniques and their Merit",
abstract = "Latency is increasingly becoming a performance bottleneck for Internet Protocol (IP) networks, but historically networks have been designed with aims of maximizing throughput and utilization. This article offers a broad survey of techniques aimed at tackling latency in the literature up to March 2014 and their merits. A goal of this work is to be able to quantify and compare the merits of the different Internet latency reducing techniques, contrasting their gains in delay reduction versus the pain required to implement and deploy them. We found that classifying techniques according to the sources of delay they alleviate provided the best insight into the following issues: 1) the structural arrangement of a network, such as placement of servers and suboptimal routes, can contribute significantly to latency; 2) each interaction between communicating endpoints adds a Round Trip Time (RTT) to latency, especially significant for short flows; 3) in addition to base propagation delay, several sources of delay accumulate along transmission paths, today intermittently dominated by queuing delays; 4) it takes time to sense and use available capacity, with overuse inflicting latency on other flows sharing the capacity; and 5) within end systems delay sources include operating system buffering, head-of-line blocking, and hardware interaction. No single source of delay dominates in all cases, and many of these sources are spasmodic and highly variable. Solutions addressing these sources often both reduce the overall latency and make it more predictable.",
author = "Bob Briscoe and Anna Brunstrom and Andreas Petlund and David Hayes and David Ross and Ing-Jyh Tsan and Stein Gjessing and Gorry Fairhurst and Carsten Griwodz and Michael Welzl",
note = "Bob Briscoe, Anna Brunstrom, Andreas Petlund, David Hayes, David Ros, Ing-Jyh Tsang, Stein Gjessing, Gorry Fairhurst, Carsten Griwodz, Michael Welzl",
year = "2014",
month = "11",
day = "26",
doi = "10.1109/COMST.2014.2375213",
language = "English",
volume = "18",
pages = "2149 -- 2196",
journal = "IEEE Communications Surveys & Tutorials",
issn = "1553-877X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "3",

}

TY - JOUR

T1 - Reducing Internet Latency

T2 - A Survey of Techniques and their Merit

AU - Briscoe, Bob

AU - Brunstrom, Anna

AU - Petlund, Andreas

AU - Hayes, David

AU - Ross, David

AU - Tsan, Ing-Jyh

AU - Gjessing, Stein

AU - Fairhurst, Gorry

AU - Griwodz, Carsten

AU - Welzl, Michael

N1 - Bob Briscoe, Anna Brunstrom, Andreas Petlund, David Hayes, David Ros, Ing-Jyh Tsang, Stein Gjessing, Gorry Fairhurst, Carsten Griwodz, Michael Welzl

PY - 2014/11/26

Y1 - 2014/11/26

N2 - Latency is increasingly becoming a performance bottleneck for Internet Protocol (IP) networks, but historically networks have been designed with aims of maximizing throughput and utilization. This article offers a broad survey of techniques aimed at tackling latency in the literature up to March 2014 and their merits. A goal of this work is to be able to quantify and compare the merits of the different Internet latency reducing techniques, contrasting their gains in delay reduction versus the pain required to implement and deploy them. We found that classifying techniques according to the sources of delay they alleviate provided the best insight into the following issues: 1) the structural arrangement of a network, such as placement of servers and suboptimal routes, can contribute significantly to latency; 2) each interaction between communicating endpoints adds a Round Trip Time (RTT) to latency, especially significant for short flows; 3) in addition to base propagation delay, several sources of delay accumulate along transmission paths, today intermittently dominated by queuing delays; 4) it takes time to sense and use available capacity, with overuse inflicting latency on other flows sharing the capacity; and 5) within end systems delay sources include operating system buffering, head-of-line blocking, and hardware interaction. No single source of delay dominates in all cases, and many of these sources are spasmodic and highly variable. Solutions addressing these sources often both reduce the overall latency and make it more predictable.

AB - Latency is increasingly becoming a performance bottleneck for Internet Protocol (IP) networks, but historically networks have been designed with aims of maximizing throughput and utilization. This article offers a broad survey of techniques aimed at tackling latency in the literature up to March 2014 and their merits. A goal of this work is to be able to quantify and compare the merits of the different Internet latency reducing techniques, contrasting their gains in delay reduction versus the pain required to implement and deploy them. We found that classifying techniques according to the sources of delay they alleviate provided the best insight into the following issues: 1) the structural arrangement of a network, such as placement of servers and suboptimal routes, can contribute significantly to latency; 2) each interaction between communicating endpoints adds a Round Trip Time (RTT) to latency, especially significant for short flows; 3) in addition to base propagation delay, several sources of delay accumulate along transmission paths, today intermittently dominated by queuing delays; 4) it takes time to sense and use available capacity, with overuse inflicting latency on other flows sharing the capacity; and 5) within end systems delay sources include operating system buffering, head-of-line blocking, and hardware interaction. No single source of delay dominates in all cases, and many of these sources are spasmodic and highly variable. Solutions addressing these sources often both reduce the overall latency and make it more predictable.

UR - http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6967689

U2 - 10.1109/COMST.2014.2375213

DO - 10.1109/COMST.2014.2375213

M3 - Article

VL - 18

SP - 2149

EP - 2196

JO - IEEE Communications Surveys & Tutorials

JF - IEEE Communications Surveys & Tutorials

SN - 1553-877X

IS - 3

ER -