An Overview on Application of Machine Learning Techniques in Optical Networks

Abstract

Today's telecommunication networks have become sources of enormous amounts of widely heterogeneous data. This information can be retrieved from network traffic traces, network alarms, signal quality indicators, users' behavioral data, etc. Advanced mathematical tools are required to extract meaningful information from these data and take decisions pertaining to the proper functioning of the networks from the network-generated data. Among these mathematical tools, Machine Learning (ML) is regarded as one of the most promising methodological approaches to perform network-data analysis and enable automated network self-configuration and fault management.The adoption of ML techniques in the field of optical communication networks is motivated by the unprecedented growth of network complexity faced by optical networks in the last few years. Such complexity increase is due to the introduction of a huge number of adjustable and interdependent system parameters (e.g., routing configurations, modulation format, symbol rate, coding schemes, etc.) that are enabled by the usage of coherent transmission/reception technologies, advanced digital signal processing and compensation of nonlinear effects in optical fiber propagation.In this paper we provide an overview of the application of ML to optical communications and networking. We classify and survey relevant literature dealing with the topic, and we also provide an introductory tutorial on ML for researchers and practitioners interested in this field. Although a good number of research papers have recently appeared, the application of ML to optical networks is still in its infancy: to stimulate further work in this area, we conclude the paper proposing new possible research directions.

A Very Brief Introduction to Machine Learning With Applications to Communication Systems

“…Considering a software-defined networking implementation, routing requires the availability at a network controller of information regarding the quality of individual communication links in the core network, as well as regarding the status of the queues at the network routers. In the presence of wireless or optical communications, the quality of a link may not be available at the network controller, but it may be predicted using available historical data [33], [54] in the absence of agreed-upon dynamic availability models. In a similar manner, predicting congestion can be framed as a data-aided classification problem [55].…”

Hop distance–based bandwidth allocation technique for elastic optical networks

“…Similarly to the training phase of machine learning algorithms, Algorithm 2 searches for good solutions to the and parameters via simulation.…”

“…Our suggestion to this problem is to diverge to a good solution by pre-processing. Similarly to the training phase of machine learning algorithms, 11,12 Algorithm 2 searches for good solutions to the activate and masks parameters via simulation.…”

Fiber Nonlinearity Mitigation via the Parzen Window Classifier for Dispersion Managed and Unmanaged Links

“…Machine learning techniques have been investigated for different applications in optical fiber communication systems [1][2][3]. Among these applications, the mitigation of fiber nonlinearity, which is the major challenge limiting the information-carrying capacity of long haul optical fiber transmission, has attracted considerable interest.…”

“…In the presence of fiber nonlinearity, the noise distribution is no longer circularly symmetric. Therefore, the optimum symbol detection requires knowledge and full parameterization of the likelihood function [1].…”