Preprocessing for segment routing optimization

In this article we introduce a preprocessing technique to solve the Segment Routing Traffic Engineering Problem optimally using significantly fewer computational resources than previously introduced methods. Segment routing is a recently developed interior gateway routing protocol to be used on top...

Full description

Saved in:
Bibliographic Details
Published in:Networks Vol. 82; no. 4; pp. 459 - 478
Main Authors: Callebaut, Hugo, De Boeck, Jérôme, Fortz, Bernard
Format: Journal Article
Language:English
Published: New York Wiley Subscription Services, Inc 01.12.2023
Wiley
Subjects:
Algorithms
Business & economic sciences
Computer networks
Computer Science
Dynamic programming
Engineering, computing & technology
Graph Theory
Ingénierie, informatique & technologie
Mixed Integer Linear Programming
Méthodes quantitatives en économie & gestion
Network Flows
Network Optimisation
Network topologies
Operations Research
Optimization
Preprocessing
Quantitative methods in economics & management
Routing Algorithms
Sciences informatiques
Sciences économiques & de gestion
Segment Routing
Segments
Solution space
Telecommunication Networks
Traffic control
Traffic Engineering
ISSN:0028-3045, 1097-0037, 1097-0037
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this article we introduce a preprocessing technique to solve the Segment Routing Traffic Engineering Problem optimally using significantly fewer computational resources than previously introduced methods. Segment routing is a recently developed interior gateway routing protocol to be used on top of existing protocols that introduces more flexibility in traffic engineering. In practice, segment routing allows to deviate traffic from its original path by specifying a list of intermediate nodes or links, called segments, to visit before going to its destination. The issue we tackle in this article is that the number of segment paths scales exponentially with the maximum number of segments allowed leading to scalability issues in mathematical formulations. This article introduces the notion of dominated segment paths, these are paths that can be eliminated from the solution space when searching for an optimal solution. We propose a dynamic programming algorithm eliminating dominated paths for any number of segments. Numerical results show that respectively 50%, 90%, and 97% of paths are dominated when considering up to 2, 3, and 4 segments on benchmark network topologies.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
scopus-id:2-s2.0-85165605380
ISSN:0028-3045
1097-0037
1097-0037
DOI:10.1002/net.22165