Spectral processing techniques for efficient monitoring in optical networks

Having ubiquitous optical monitors in dense wavelength-division multiplexing (DWDM) or flex-grid networks allows the estimation real time of crucial parameters. Such monitoring would be even more important disaggregated networks, to inspect performance issues related inter-vendor interoperability. Several parameters can retrieved using spectrum analyzers (OSAs). However, omnipresent OSAs represent an infeasible solution. Nevertheless, advent new, relatively cheap, compact and medium-resolution channel (OCMs) enable a intensive deployment these devices. In this paper, we identify two main scenarios for placement such monitors: at ingress egress nodes. scenario, directly estimate signals, but not those filters. On contrary, filter-related easily detected, amplified spontaneous emission. Therefore, present methods that, leveraging curve fitting machine learning regression algorithm, allow detection missing We verify proposed solutions with spectral data acquired simulation experimental setups. obtained good accuracy both setups studied scenarios. It is noteworthy that assessment achieved maximum absolute error (MAE) lower than 1 GHz filter bandwidth MAE 0.5 frequency shift estimation. addition, by comparing relative errors considered parameters, identified scenario as beneficial. particular, estimated central 84% 6 dB 75% higher accuracy, respect datasheet/reference values. This translates into total reduction signal-to-noise ratio (SNR) penalty, introduced single filter, 0.24 dB.