Analysis of OSNR and Data Rate Selection Using ML Techniques for Optical Networks

Abstract

Optical networks are preferred over other wireless network in modern world due to better link quality in long haul networks. The basic parameters that define any optical network is Quality factor, OSNR, BER etc. As the length of optical fiber is increased then link quality is diminished due to attenuation, PMD, and various other non nonlinear factors. Our aim of this paper is to compute the path metric for Optical OFDM network with the higher data rates as 100Gbps, 200Gbps, 300Gbps, 400Gbps and 500Gbps and with varying the fiber length. Hence, to take all non-linearity and attenuation effect into account and see how our system behaves. We employed fiber length from 1km to 130km with attenuation constant of 0.2dB/km. By the incorporation of machine learning cognition is introduced in network control plane. Our main idea is to make intelligent network layer by means of this Adaptive Neural Fuzzy Inference System such that any path with any data rate demanding for connection in any single node network. Based on their path metric i.e. OSNR and performance of specific data rate for that node to node length the network will pass the data with required data rate, or due to poor performance in the require data rate lesser data rate path is assigned.