Reach-related Energy Consumption in IP-over-WDM 100G Translucent Networks

In this paper we investigate the power consumption of IP-over-WDM optical networks as a function of 100 Gb/s Transponders' (TXPs) transmission reach when the "All-On and the "On-Off" planning approaches are considered. The former considers that transmission devices are kept always on and provides an upper bound to the network power consumption whereas the latter maximizes power savings by switching devices on-off according to daily traffic variations. For both approaches we evaluate the impact of the optical reach when a) two different Multi-Layer design strategies are used, b) the demanded capacity per traffic request increases, and c) 100G TXPs are subject to a future possible power consumption reduction. This analysis has been carried out by using a novel analytical model that captures the increase of TXP power consumption with optical reach. Simulations on an European-like backbone network show that there exists an optical reach that leads to a minimum network power consumption (i.e., optimal reach). Energy savings brought by the optimal reach depend on the total aggregated traffic, the traffic variation during the day, the design strategy and the ratio of the processing costs in terms of Watt/Gb/s between IP and WDM layer. However, the optical reach which guarantees the lowest average consumption during the day in the On-Off planning corresponds to the optimal reach in the All-On scenario. Finally, we show that the benefits of a Multi-Layer design strategy which enables traffic grooming decrease when WDM-layer power consumption decreases and when the demanded capacity per traffic request increases, both for the All-On and the On-Off planning.