An Experimental Study of the Learnability of Congestion Control

When designing a distributed network protocol, typically it is infeasible to fully define the target network where the protocol is intended to be used. It is therefore natural to ask: How faithfully do protocol designers really need to understand the networks they design for? What are the important...

Full description

Bibliographic Details
Main Authors: Sivaraman Kaushalram, Anirudh, Winstein, Keith, Thaker, Pratiksha R., Balakrishnan, Hari
Other Authors: Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
Format: Article
Language:en_US
Published: Association for Computing Machinery (ACM) 2014
Online Access:http://hdl.handle.net/1721.1/88914
https://orcid.org/0000-0002-1455-9652
https://orcid.org/0000-0003-4034-0918
_version_ 1811082225235525632
author Sivaraman Kaushalram, Anirudh
Winstein, Keith
Thaker, Pratiksha R.
Balakrishnan, Hari
author2 Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
author_facet Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
Sivaraman Kaushalram, Anirudh
Winstein, Keith
Thaker, Pratiksha R.
Balakrishnan, Hari
author_sort Sivaraman Kaushalram, Anirudh
collection MIT
description When designing a distributed network protocol, typically it is infeasible to fully define the target network where the protocol is intended to be used. It is therefore natural to ask: How faithfully do protocol designers really need to understand the networks they design for? What are the important signals that endpoints should listen to? How can researchers gain confidence that systems that work well on well-characterized test networks during development will also perform adequately on real networks that are inevitably more complex, or future networks yet to be developed? Is there a tradeoff between the performance of a protocol and the breadth of its intended operating range of networks? What is the cost of playing fairly with cross-traffic that is governed by another protocol? We examine these questions quantitatively in the context of congestion control, by using an automated protocol-design tool to approximate the best possible congestion-control scheme given imperfect prior knowledge about the network. We found only weak evidence of a tradeoff between operating range in link speeds and performance, even when the operating range was extended to cover a thousand-fold range of link speeds. We found that it may be acceptable to simplify some characteristics of the network—such as its topology—when modeling for design purposes. Some other features, such as the degree of multiplexing and the aggressiveness of contending endpoints, are important to capture in a model.
first_indexed 2024-09-23T11:59:42Z
format Article
id mit-1721.1/88914
institution Massachusetts Institute of Technology
language en_US
last_indexed 2024-09-23T11:59:42Z
publishDate 2014
publisher Association for Computing Machinery (ACM)
record_format dspace
spelling mit-1721.1/889142022-09-27T23:19:39Z An Experimental Study of the Learnability of Congestion Control Sivaraman Kaushalram, Anirudh Winstein, Keith Thaker, Pratiksha R. Balakrishnan, Hari Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Balakrishnan, Hari Sivaraman Kaushalram, Anirudh Winstein, Keith Thaker, Pratiksha R. Balakrishnan, Hari When designing a distributed network protocol, typically it is infeasible to fully define the target network where the protocol is intended to be used. It is therefore natural to ask: How faithfully do protocol designers really need to understand the networks they design for? What are the important signals that endpoints should listen to? How can researchers gain confidence that systems that work well on well-characterized test networks during development will also perform adequately on real networks that are inevitably more complex, or future networks yet to be developed? Is there a tradeoff between the performance of a protocol and the breadth of its intended operating range of networks? What is the cost of playing fairly with cross-traffic that is governed by another protocol? We examine these questions quantitatively in the context of congestion control, by using an automated protocol-design tool to approximate the best possible congestion-control scheme given imperfect prior knowledge about the network. We found only weak evidence of a tradeoff between operating range in link speeds and performance, even when the operating range was extended to cover a thousand-fold range of link speeds. We found that it may be acceptable to simplify some characteristics of the network—such as its topology—when modeling for design purposes. Some other features, such as the degree of multiplexing and the aggressiveness of contending endpoints, are important to capture in a model. National Science Foundation (U.S.) (Grant CNS-1040072) 2014-08-19T18:46:14Z 2014-08-19T18:46:14Z 2014-08 Article http://purl.org/eprint/type/ConferencePaper 978-1-4503-2836-4 http://hdl.handle.net/1721.1/88914 Sivaraman, Anirudh, Keith Winstein, Pratiksha Thanker, and Hari Balakrishnan. "An Experimental Study of the Learnability of Congestion Control." Proceedings of the 2014 ACM conference on SIGCOMM, August 17-22, 2014, Chicago, IL. https://orcid.org/0000-0002-1455-9652 https://orcid.org/0000-0003-4034-0918 en_US http://cs.stanford.edu/~keithw/www/Learnability-SIGCOMM2014.pdf Proceedings of the 2014 ACM conference on SIGCOMM Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Association for Computing Machinery (ACM) Sivaraman
spellingShingle Sivaraman Kaushalram, Anirudh
Winstein, Keith
Thaker, Pratiksha R.
Balakrishnan, Hari
An Experimental Study of the Learnability of Congestion Control
title An Experimental Study of the Learnability of Congestion Control
title_full An Experimental Study of the Learnability of Congestion Control
title_fullStr An Experimental Study of the Learnability of Congestion Control
title_full_unstemmed An Experimental Study of the Learnability of Congestion Control
title_short An Experimental Study of the Learnability of Congestion Control
title_sort experimental study of the learnability of congestion control
url http://hdl.handle.net/1721.1/88914
https://orcid.org/0000-0002-1455-9652
https://orcid.org/0000-0003-4034-0918
work_keys_str_mv AT sivaramankaushalramanirudh anexperimentalstudyofthelearnabilityofcongestioncontrol
AT winsteinkeith anexperimentalstudyofthelearnabilityofcongestioncontrol
AT thakerpratikshar anexperimentalstudyofthelearnabilityofcongestioncontrol
AT balakrishnanhari anexperimentalstudyofthelearnabilityofcongestioncontrol
AT sivaramankaushalramanirudh experimentalstudyofthelearnabilityofcongestioncontrol
AT winsteinkeith experimentalstudyofthelearnabilityofcongestioncontrol
AT thakerpratikshar experimentalstudyofthelearnabilityofcongestioncontrol
AT balakrishnanhari experimentalstudyofthelearnabilityofcongestioncontrol