All-optical narrowband spectral slicing encryption with super-Gaussian filters

Recent literature points out the benefits of encrypting signals in the physical layer to improve network security. In particular, a novel all-optical encryption approach uses narrowband optical band-pass filters (OBPF) to split a dense wavelength division multiplexing (DWDM)-compatible signal into several spectral slices which, then, have their physical properties, such as phase and delay, altered; as a result, after multiplexing all slices, an encoded version of the input signal is obtained. The performance of such technique was previously evaluated by assuming that all OBPFs are characterized by ideal rectangular-shaped transfer functions. In this work, we investigate how such performance is affected by the utilization of practical OBPFs with super-Gaussian profiles. Simulation results indicate that there is a trade-off between the filter bandwidth and filter order that may allow for encrypted signals to be properly decoded even after propagation distances up to 400 km, which is typical metropolitan area network distance.