Advanced

New Results on Multipath Routing

Pioro, Michal LU ; Mohanraj, Sudhir; Medhi, Deep and Liu, Xuan (2014) 2014 INFORMS Telecommunications Conference
Abstract
Multipath routing of traffic demands, known also as bifurcated routing or load sharing,

has been studied for various telecommunication networks for a long time. An advantage of

multipath routing is that it gives a demand the opportunity to use multiple paths through

the network to send its traffic. In a single-commodity situation, this benefit can be clearly

seen. When multiple demands (multi-commodity situation) compete for the same resources

in a network, we focus on the problem of how many demands can take real advantage of

multiple paths. In this context, we present new results on multipath routing for a number

of network traffic objectives. We show that under certain... (More)
Multipath routing of traffic demands, known also as bifurcated routing or load sharing,

has been studied for various telecommunication networks for a long time. An advantage of

multipath routing is that it gives a demand the opportunity to use multiple paths through

the network to send its traffic. In a single-commodity situation, this benefit can be clearly

seen. When multiple demands (multi-commodity situation) compete for the same resources

in a network, we focus on the problem of how many demands can take real advantage of

multiple paths. In this context, we present new results on multipath routing for a number

of network traffic objectives. We show that under certain traffic conditions and topological

structures, multipath routing provides virtually no gain compared to single-path routing

when the traffic is offered for all demand pairs in a network. We also present results on how

different network objectives influence the ability of taking advantage of multipath routing.

Our results, based on the basic properties of linear programming, are somewhat against

a rather common belief (expressed by the term ”load sharing”) that multipath routing is

significantly more effective in carrying traffic than single-path routing. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to conference
publication status
published
subject
conference name
2014 INFORMS Telecommunications Conference
language
English
LU publication?
yes
id
00193d33-27be-4179-91a8-5a1379407cf2 (old id 5337448)
date added to LUP
2015-04-29 14:25:16
date last changed
2016-04-16 10:50:49
@misc{00193d33-27be-4179-91a8-5a1379407cf2,
  abstract     = {Multipath routing of traffic demands, known also as bifurcated routing or load sharing,<br/><br>
has been studied for various telecommunication networks for a long time. An advantage of<br/><br>
multipath routing is that it gives a demand the opportunity to use multiple paths through<br/><br>
the network to send its traffic. In a single-commodity situation, this benefit can be clearly<br/><br>
seen. When multiple demands (multi-commodity situation) compete for the same resources<br/><br>
in a network, we focus on the problem of how many demands can take real advantage of<br/><br>
multiple paths. In this context, we present new results on multipath routing for a number<br/><br>
of network traffic objectives. We show that under certain traffic conditions and topological<br/><br>
structures, multipath routing provides virtually no gain compared to single-path routing<br/><br>
when the traffic is offered for all demand pairs in a network. We also present results on how<br/><br>
different network objectives influence the ability of taking advantage of multipath routing.<br/><br>
Our results, based on the basic properties of linear programming, are somewhat against<br/><br>
a rather common belief (expressed by the term ”load sharing”) that multipath routing is<br/><br>
significantly more effective in carrying traffic than single-path routing.},
  author       = {Pioro, Michal and Mohanraj, Sudhir and Medhi, Deep and Liu, Xuan},
  language     = {eng},
  title        = {New Results on Multipath Routing},
  year         = {2014},
}