Skip to Main Content
In order to mitigate the greenhouse effects and reduce environmental pollution, energy saving has become an important issue in designing the next generation Internet. Shutting down the network devices carrying light loads and redirecting their traffic flows to other routes is a way to decrease energy consumption. An energy efficient network has to dynamically determine the optimal active links to adapt itself to network traffic changes. However, in current IP networks, shutting down and/or turning on links would trigger link state routing protocols to reconverge to a new topology. Since the convergence time would take tens of seconds, routing table inconsistencies among routers would result in network disconnection and even worse, generating traffic loops during the convergence interval. In this paper, we propose a comprehensive solution to resolve such problems. The contribution of the paper is presented in three parts. First, we propose a distributed energy-aware link management algorithm to dynamically determine the link states for a router. Routers performing the algorithm obtain a pair of high and low thresholds in order to compare flow amount to determine the link states. Those thresholds are automatically adapted to network traffic so that network load and power saving are balanced. The second contribution of the paper comes from the flows being immediately redirected to their new next hop nodes when a link's state is changed. We borrow link metric from link state routing protocols to determine potential for each node pair. The potential is used to determine loop free next hop routing nodes. This technique enables us to redirect flows on the fly without suffering from long convergence of link state routing protocols. The third contribution of the work comes from our proposal of an optimization model to determine a link metric and a most power saving network topology. The obtained link metric is used in our distributed link management algorithm to determine - ode potential. We have evaluated the performance of the proposed algorithm through extensive simulations. The numerical results reveal that the proposed distributed link management algorithm can effectively save network energy in three benchmark networks.