We study the possibility of bypassing the table lookup capability of router linecards to save energy in IP networks. Differently from previous works, the table lookup operation is bypassed independently for each linecard. We first present a model for our scenario. Then, the problem is formally stated as an optimization problem. Moreover, we provide a genetic algorithm to solve the problem up to medium sized network. We evaluate the performance of the optimal solution and the genetic algorithm on synthetically generated topologies. We then consider large topologies available in the literature to evaluate the performance of our algorithm. We compare the table lookup operation with the traditional operation of switching off the linecards. Results show that in general the table lookup operation achieve higher energy savings than the simple switch off. Then, we consider the impact of jointly applying the two techniques to further increase the energy savings. In this case, we show that the saving can be further increased by application of the two techniques. (C) 2014 Elsevier B.V. All fights reserved..
Reducing power consumption in backbone IP networks through table lookup bypass / Coiro, Angelo; Chiaraviglio, Luca; Cianfrani, Antonio; Listanti, Marco; Polverini, Marco. - In: COMPUTER NETWORKS. - ISSN 1389-1286. - 64:(2014), pp. 125-142. [10.1016/j.comnet.2014.02.005]
Reducing power consumption in backbone IP networks through table lookup bypass
COIRO, ANGELO;CHIARAVIGLIO, LUCA;CIANFRANI, Antonio;LISTANTI, Marco;POLVERINI, MARCO
2014
Abstract
We study the possibility of bypassing the table lookup capability of router linecards to save energy in IP networks. Differently from previous works, the table lookup operation is bypassed independently for each linecard. We first present a model for our scenario. Then, the problem is formally stated as an optimization problem. Moreover, we provide a genetic algorithm to solve the problem up to medium sized network. We evaluate the performance of the optimal solution and the genetic algorithm on synthetically generated topologies. We then consider large topologies available in the literature to evaluate the performance of our algorithm. We compare the table lookup operation with the traditional operation of switching off the linecards. Results show that in general the table lookup operation achieve higher energy savings than the simple switch off. Then, we consider the impact of jointly applying the two techniques to further increase the energy savings. In this case, we show that the saving can be further increased by application of the two techniques. (C) 2014 Elsevier B.V. All fights reserved..I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.