In this paper new load balancing techniques are proposed based on IP-like routing, where the forwarding mechanism is only driven by the destination address. The aim of this work is to consider the applicability of this approach in the context of optical networks. The proposed algorithm (RSNE - Reverse Subtree Neighborhood Exploration) implements a local search technique where the basic step is the modification of a single entry in the routing table of a node. A randomized method with reduced computational complexity (fRSNE - fast RSNE) is also presented and analyzed. Both schemes allow an incremental implementation where local search steps are continuously performed as traffic conditions change. Experiments under static and dynamic (time varying) traÆc scenarios show a rapid reduction of the congestion. The performance of the incremental scheme while tracking a changing traffic matrix is comparable to that obtainable through the complete re-optimization of the traffic, while the randomized implementation is particularly efficient when scaling properties are considered.
Dynamic load balancing in WDM Networks / Brunato, Mauro; Battiti, Roberto; Salvadori, Elio. - ELETTRONICO. - (2002), pp. 1-21.
Dynamic load balancing in WDM Networks
Brunato, Mauro;Battiti, Roberto;Salvadori, Elio
2002-01-01
Abstract
In this paper new load balancing techniques are proposed based on IP-like routing, where the forwarding mechanism is only driven by the destination address. The aim of this work is to consider the applicability of this approach in the context of optical networks. The proposed algorithm (RSNE - Reverse Subtree Neighborhood Exploration) implements a local search technique where the basic step is the modification of a single entry in the routing table of a node. A randomized method with reduced computational complexity (fRSNE - fast RSNE) is also presented and analyzed. Both schemes allow an incremental implementation where local search steps are continuously performed as traffic conditions change. Experiments under static and dynamic (time varying) traÆc scenarios show a rapid reduction of the congestion. The performance of the incremental scheme while tracking a changing traffic matrix is comparable to that obtainable through the complete re-optimization of the traffic, while the randomized implementation is particularly efficient when scaling properties are considered.File | Dimensione | Formato | |
---|---|---|---|
11.pdf
accesso aperto
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
472.28 kB
Formato
Adobe PDF
|
472.28 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione