Performance of translucent optical networks under dynamic traffic and uncertain physical-layer information

This paper investigates the performance of translucent optical transport networks (OTNs) under different traffic and knowledge conditions, varying from perfect knowledge to drifts and uncertainties in the physical-layer parameters. Our focus is on the regular operation of a translucent OTN, i.e., after the dimensioning and regenerator placement phase. Our contributions can be summarized as follows. Based on the computation of the Personicks Q factor, we introduce a new methodology for the assessment of the optical signal quality along a path, and show its application on a realistic example. We analyze the performance of both deterministic and predictive RWA techniques integrating this signal quality factor Q in the lightpath computation process. Our results confirm the effectiveness of predictive techniques to deal with the typical drifts and uncertainties in the physical-layer parameters, in contrast to the superior efficacy of deterministic approaches in case of perfect knowledge. Conversely to most previous works, where all wavelengths are assumed to have the same characteristics, we examine the case when the network is not perfectly compensated, so the Maximum Transmission Distance (MTD) of the different wavelength channels may vary. We show that blocking might increase dramatically when the MTD of the different wavelength channels is overlooked.