Using Topology Aggregation for Efficient Shared Segment Protection Solutions in Multi-Domain Networks

The dynamic routing problem for Overlapping Segment Shared Protection (OSSP) in multi-domain networks has not received a lot of interest so far as it is more complex than in single-domain networks. Difficulties lie in the lack of complete and global knowledge about network topologies and bandwidth a...

Full description

Saved in:
Bibliographic Details
Published in:IEEE journal on selected areas in communications 2007-12, Vol.25 (9), p.96-107
Main Authors: Dieu-Linh Truong, Jaumard, B.
Format: Article
Language:English
Subjects:
Agglomeration
Bandwidth
Channel allocation
Computer networks
Computer science
Costs
Heuristic
Information systems
Mathematical models
Network topologies
Network topology
Networks
Operations research
Optimization
Protection
Routing
Routing (telecommunications)
Segments
Studies
Topology
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The dynamic routing problem for Overlapping Segment Shared Protection (OSSP) in multi-domain networks has not received a lot of interest so far as it is more complex than in single-domain networks. Difficulties lie in the lack of complete and global knowledge about network topologies and bandwidth allocation whereas this knowledge is easily available in single-domain networks. We propose a two-step routing approach for the OSSP based on a topology aggregation scheme and link cost estimation: an inter-domain step and an intra-domain step. We propose two different heuristics, GROS and DYPOS for the inter-domain step, and a "Blocking-go-back" strategy in order to reduce the blocking rate in the intra-domain step. We compare the performance of the two heuristics against an optimal single-domain approach. We show that both heuristics lead to resource efficient solutions that are not far from the optimal ones. Moreover, both heuristics require relatively small computational efforts and are scalable for multi-domain networks.
ISSN:0733-8716
1558-0008
DOI:10.1109/JSAC-OCN.2007.028606