Green grooming in spectrum-sliced elastic optical path networks

With the continuous expansion in network scale and the rapid growth of Internet traffic, a high-capacity and power-efficient transport system is required. The Spectrum-Sliced Elastic Optical Path Network (SLICE) based on Orthogonal Frequency Division Multiplexing has been considered as a promising solution due to flexible spectrum provisioning. In a SLICE, compared with traditional components, the bandwidth variable transponder and cross-connects have higher power costs. So it is necessary to make a green footprint in this new networking framework. In this paper, we make a comprehensive study of green grooming in SLICEs, with the objective to save both spectrum and power consumptions. An Integer Linear Programming formulation with various SLICE constraints is presented for the optimization problem above. Under the large-scale network scenario, we also propose the auxiliary graph model which reflects spectrum/power usage, grooming policies implemented by adjusting link costs according to a comprehensive power consumption model, as well as a threshold-based spectrum reservation scheme. As a result, an efficient heuristic called green grooming algorithm is designed. The extensive simulation results demonstrate that our heuristic obtains near-optimal results and achieves a better performance in terms of spectrum and power efficiencies compared with benchmarks, under various topology structures.