Survey on Machine Learning for Traffic-Driven Service Provisioning in Optical Networks

The unprecedented growth of the global Internet traffic, coupled with large spatio-temporal fluctuations that create, to some extent, predictable tidal traffic conditions, are motivating evolution from reactive proactive and eventually towards adaptive optical networks. In these networks, traffic-driven service provisioning can address problem network over-provisioning better adapt variations, while keeping quality-of-service at required levels. Such an approach will reduce resource thus total cost. This survey provides a comprehensive review state art on machine learning (ML)-based techniques layer for provisioning. in networks is initially presented, followed by overview ML utilized ML-aided approaches presented detail, including predictive prescriptive frameworks For all outlined, discussion their limitations, research challenges, potential opportunities also presented.