Topology-aware approach for reducing power consumption in backbone networks

According to several studies, line cards are the major source of energy savings in green networks because they contribute up to 42% of the total energy consumption of a backbone router. Since these networks are over provisioned, during periods of low-demands shutting down network links seems like a good approach to reduce energy consumption. In this paper, we propose a novel topology-aware approach to exploit the sleep mode capabilities of links in IP networks. We leverage the algebraic connectivity metric, adopted from spectral graph theory on IP networks in order to identify the set of links, that can be powered off by decreasing the network connectivity the least. Exhaustive searching for the optimal link shutting down is computationally infeasible. Hence, we propose several strategies that select single link for removal, based on topological and spectral metrics. Moreover, the proposed algorithms do not require the knowledge of traffic matrix, being able to run in real-time, where this information would not be available. Results, obtained on real topologies prove that our algorithms are able to achieve performance comparable to a pioneering more complex traffic aware approaches, called benchmark (Zhang et al., 2010) and least-flow (Chiaraviglio et al., 2009).