Energy Aware Resource Allocation Based on Demand Bandwidth and Duration

AbstractThe tremendous growth in high-bandwidth applications and consequent increase in power consumption has underscored the importance of energy efficient design strategies for backbone optical networks. In this paper, we first present an ef–ficient integer linear program (ILP) formulation for allocating resources to a sub-wavelength traffic demand, such that the increase in energy consumption due to the new demand is minimized. The ILP allocates resources to the new de–mand, based on the bandwidth usage and durations of existing demands, as well as that of the new demand. We then extend this approach, and propose a heuristic that performs resource allocation for a set of scheduled, periodic demands with the goal of reducing the overall energy consumption of the network for the entire demand set. We show through simulations that the proposed ILP can quickly generate solutions for practical networks, with hundreds of demands and demonstrate that knowledge of the demand holding times can be exploited to significantly reduce the overall network energy consumption