Larroca Rougier 2009a - Revision history

Scipediacontent at 12:47, 21 January 2021

2021-01-21T12:47:53Z

Scipediacontent: Scipediacontent moved page Draft Content 409150014 to Larroca Rougier 2009a

2020-10-14T14:02:40Z

Scipediacontent moved page Draft Content 409150014 to Larroca Rougier 2009a

Scipediacontent: Created page with " == Abstract == Network convergence and new applications running on end-hosts result in increasingly variable and unpredictable traffic patterns. By providing origin-destinat..."

2020-10-14T14:02:37Z

Created page with " == Abstract == Network convergence and new applications running on end-hosts result in increasingly variable and unpredictable traffic patterns. By providing origin-destinat..."

New page

== Abstract ==

Network convergence and new applications running on end-hosts result in increasingly variable and unpredictable traffic patterns. By providing origin-destination pairs with several possible paths, load-balancing has proved itself an excellent tool to face this uncertainty. Formally, load-balancing is defined in terms of a convex link cost function of its load, where the objective is to minimize the total cost. Typically, the link queueing delay is used as this cost since it measures its congestion. Over-simplistic models are used to calculate it, which have been observed to result in suboptimal resource usage and total delay. In this paper we investigate the possibility of learning the delay function from measurements, thus converging to the actual minimum. A novel regression method is used to make the estimation, restricting the assumptions to the minimum (e.g. delay should increase with load). The framework is relatively simple to implement, and we discuss some possible variants.

Document type: Part of book or chapter of book

== Full document ==
<pdf>Media:Draft_Content_409150014-beopen921-5072-document.pdf</pdf>

== Original document ==

The different versions of the original document can be found in:

* [https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf]

← Older revision		Revision as of 12:47, 21 January 2021
Line 3:		Line 3:

	Network convergence and new applications running on end-hosts result in increasingly variable and unpredictable traffic patterns. By providing origin-destination pairs with several possible paths, load-balancing has proved itself an excellent tool to face this uncertainty. Formally, load-balancing is defined in terms of a convex link cost function of its load, where the objective is to minimize the total cost. Typically, the link queueing delay is used as this cost since it measures its congestion. Over-simplistic models are used to calculate it, which have been observed to result in suboptimal resource usage and total delay. In this paper we investigate the possibility of learning the delay function from measurements, thus converging to the actual minimum. A novel regression method is used to make the estimation, restricting the assumptions to the minimum (e.g. delay should increase with load). The framework is relatively simple to implement, and we discuss some possible variants.		Network convergence and new applications running on end-hosts result in increasingly variable and unpredictable traffic patterns. By providing origin-destination pairs with several possible paths, load-balancing has proved itself an excellent tool to face this uncertainty. Formally, load-balancing is defined in terms of a convex link cost function of its load, where the objective is to minimize the total cost. Typically, the link queueing delay is used as this cost since it measures its congestion. Over-simplistic models are used to calculate it, which have been observed to result in suboptimal resource usage and total delay. In this paper we investigate the possibility of learning the delay function from measurements, thus converging to the actual minimum. A novel regression method is used to make the estimation, restricting the assumptions to the minimum (e.g. delay should increase with load). The framework is relatively simple to implement, and we discuss some possible variants.
−
−	~~Document type: Part of book or chapter of book~~
−
−	~~== Full document ==~~
−	~~<pdf>Media:Draft_Content_409150014-beopen921-5072-document.pdf</pdf>~~


Line 15:		Line 10:

	* [https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf]		* [https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf https://link.springer.com/content/pdf/10.1007%2F978-3-642-01399-7_61.pdf]
		+
		+	* [http://link.springer.com/content/pdf/10.1007/978-3-642-01399-7_61 http://link.springer.com/content/pdf/10.1007/978-3-642-01399-7_61],
		+	: [http://dx.doi.org/10.1007/978-3-642-01399-7_61 http://dx.doi.org/10.1007/978-3-642-01399-7_61] under the license http://www.springer.com/tdm
		+
		+	* [https://core.ac.uk/display/103696956 https://core.ac.uk/display/103696956],
		+	: [https://dl.acm.org/citation.cfm?id=1560209 https://dl.acm.org/citation.cfm?id=1560209],
		+	: [https://www.scipedia.com/public/Larroca_Rougier_2009a https://www.scipedia.com/public/Larroca_Rougier_2009a],
		+	: [http://iie.fing.edu.uy/~flarroca/papers/networking09/main.pdf http://iie.fing.edu.uy/~flarroca/papers/networking09/main.pdf],
		+	: [https://dblp.uni-trier.de/db/conf/networking/networking2009.html#LarrocaR09 https://dblp.uni-trier.de/db/conf/networking/networking2009.html#LarrocaR09],
		+	: [https://link.springer.com/chapter/10.1007/978-3-642-01399-7_61 https://link.springer.com/chapter/10.1007/978-3-642-01399-7_61],
		+	: [http://www.perso.telecom-paristech.fr/~rougierj/Pub/lr-networking09.pdf http://www.perso.telecom-paristech.fr/~rougierj/Pub/lr-networking09.pdf],
		+	: [http://biblio.telecom-paristech.fr/cgi-bin/download.cgi?id=8928 http://biblio.telecom-paristech.fr/cgi-bin/download.cgi?id=8928],
		+	: [https://academic.microsoft.com/#/detail/1547047001 https://academic.microsoft.com/#/detail/1547047001]

← Older revision	Revision as of 14:02, 14 October 2020
(No difference)