Optimal launch power prediction of a 100G PM-DQPSK dispersion-managed link with the Gaussian noise model

Abstract

Of all the non-linear fiber propagation models proposed over the years, the Gaussian Noise (GN) model is growing in popularity due to its simplicity and yet reliability when it comes to predict performance of uncompensated coherent transmission (UT) systems that rely on state-of-the art digital-signal processing (DSP) for dispersion compensation. However, many of the systems currently deployed rely on optical CD compensation. Overhauling or upgrading these systems with the most recent DSP is not always feasible. In this context, it is important to broad the range of the GNmodel to dispersion managed (DM) systems, so both scenarios can benefit from a low complexity, fast and reliable performance prediction tool. In this paper, we validate the first results comparing the performance in both accuracy and simulation time of the GN model simulating a realistic DM scenario that relies on periodical spans of non-dispersion shifted fiber (NDSF) to perform the dispersion compensation. The same realistic scenarios were modeled with commercial software and the GN model. The objective was to predict the optimal launch power for different link lengths, central wavelengths and channel spacing values. Preliminary results obtained with the GN model are in good agreement with the ones from the commercial software for several link distances tested up to 2400 Km.