Can Channel Power Optimization with GSNR Flatness Maximize Capacities of C+L-Band Optical Systems and Networks?
Ningning Guo, Gangxiang Shen, Ning Deng, Biswanath Mukherjee
Abstract
Channel power optimization for better transmission performance is essential but complex in C+L-band systems and networks due to power transfer caused by the inter-channel stimulated Raman scattering (ISRS) effect. Channel power optimization to maximize channel generalized signal-to-noise ratio (GSNR) flatness has become a common practice in industry due to its balance in channel capacity and network operational simplicity. Nonetheless, this raises an important question: can channel power optimization with GSNR flatness achieve maximum capacity for C+L-band systems and networks? For this, we focus on two optimization objectives: solely maximizing total capacity and jointly maximizing average channel GSNR and channel GSNR flatness; and we introduce a GSNR estimation model to evaluate channel signal quality and an enhanced particle swarm optimization (PSO) algorithm for channel power optimization. Moreover, to maximize total network capacity, we extend a GSNR-aware traffic grooming algorithm, in which three wavelength assignment sequences are considered. Simulation studies show that, in C+L-band systems, channel power optimization with GSNR flatness can maximize total Shannon capacity in all systems and achievable capacity in short-distance systems, which however fails to maximize achievable capacity in long-distance systems. In the context of optical networks, wavelength assignment sequences do not impact total network capacity under channel power optimization with GSNR flatness, whereas channel power optimization with maximum GSNR can maximize total network capacity under wavelength assignment sequences preferring to low-margin and high-frequency channels. Furthermore, by comparing network capacity difference under the two optimization objectives, it is observed that channel power optimization, focusing on maximizing channel average GSNR and channel GSNR flatness, can achieve a network capacity close to solely maximizing channel GSNRs in an optical network with short average link length. However, this difference grows as the average link length in the network increases. This suggests that channel power optimization with GSNR flatness is advisable for metro-area or regional networks considering its simpler network configuration. However, for long-haul networks, channel power optimization should target to maximize channel GSNRs to exploit its potential network capacity.