Fast, scalable, and distributed restoration in general mesh optical networks

Citation
Gp. Austin et al., Fast, scalable, and distributed restoration in general mesh optical networks, BELL LABS T, 6(1), 2001, pp. 67-81
Citations number
13
Categorie Soggetti
Information Tecnology & Communication Systems
Journal title
BELL LABS TECHNICAL JOURNAL
ISSN journal
10897089 → ACNP
Volume
6
Issue
1
Year of publication
2001
Pages
67 - 81
Database
ISI
SICI code
1089-7089(200101/06)6:1<67:FSADRI>2.0.ZU;2-I
Abstract
Service providers are demanding transport network solutions that can accomm odate exponential traffic growth and, at the same time, provide novel servi ces such as point-and-click provisioning of very high bandwidth circuits, o ptical bandwidth service management, fast protection and restoration, and b andwidth on demand. It is becoming increasingly clear that the required sca lability (terabits/s to petabits/s) and cost structure can only be provided by transparent optical cross connects (OXCs). The challenge, then, is to m ake the optical network consisting of OXCs, dense wavelength division multi plexers (DWDMs), and optical add/drop multiplexers (OADMs) dynamic and inte lligent. A major aspect of this intelligence is fast provisioning and resto ration. In this paper, we present a fast, scalable, and distributed solutio n for optical layer restoration in general mesh-type optical networks, whic h is being implemented as part of the Optical Navigator System (ONS) residi ng in Lucent's LambdaRouter product. The key ingredients to our solution ar e a fast and scalable restoration strategy, a fast and scalable connection setup strategy, a contention-free wavelength assignment strategy, and a fas t and reliable data communications network to exchange signaling messages. We also introduce novel concepts of demand bundling and optical virtual pat hs that ensure that restoration performance scales with network and traffic volumes. Together these components provide an intelligent optical networki ng solution that not only guarantees restoration times within few hundreds of milliseconds, but also achieves scalability.