Abstract:The evolution of optical networks is enabled by both technological and architectural advances with the goal of reducing operational and capital expenditure per transmitted bit. While the former one stimulates significant system capacity, the latter paves the way for reducing the effective traffic load in network so that more traffic can be carried. Accordingly, optical node architectures have been transitioning from optical-electrical-optical mode to all-optical one, leveraging the scalability and efficiency of fully optical cross-connecting. Conventional wisdom in designing and architecting such switching nodes is nevertheless rooted in the intuition that when an optical channel crossing an intermediate node, it should be maximally isolated from other optical channels in order to avoid interference which may result in degrading signal quality. Such long-established paradigm perceiving the interference of optical channels transiting at the same node as an adversarial factor and should therefore circumvent, albeit reasonable, may leave vast unexplored opportunities. Indeed, the rapid advances in all-optical signal processing technologies has brought opportunities to re-define the optical node architecture by upgrading its naive functionalities from simply add/drop and cross-connecting to proactively mixing optical channels in photonic domain. Specifically, all-optical channel aggregation and de-aggregation technologies have been remarkably advancing in recent years, permitting two or more optical channels at lower bit-rate and/or modulation formats could be all-optically aggregated to a single channel of higher-rate and/or higher-order modulation format and vice versa. Such evolutionary technique is poised to disrupt the existing ecosystem for optical network design and planning, and thus necessitates for a radical change to unlock new potentials. To that end, this paper presents a new paradigm for future optical networks, namely, optical-processing-enabled networks , which are powered by in-network all-optical mixing capability. We introduce the operational principle of optical channel aggregation (de-aggregation) and show how spectrally beneficial such innovative operations could yield by an illustrative example. Next, in order to maximize the aggregation opportunity, we present a mathematical model for optimal routing based on integer linear programming model. Numerical results on the realistic network topology COST239 are provided to quantify the spectral gain of aggregation-aware routing compared to the conventional one.

Evolution towards Smart Optical Networking: Where Artificial Intelligence (AI) meets the World of Photonics

Artificial intelligence (AI) methods in optical networks: A comprehensive survey

Towards Intelligent Optical Networks: The Role of Intellectual Property

Intelligent Photonics: A Disruptive Technology to Shape the Present and Redefine the Future

Study of Optical Networks, 5G, Artificial Intelligence and Their Applications

Intent-based AI system in packet-optical networks towards 6G

Toward Intelligence in Photonic Systems

Advances in Optical Communications and Network Technologies.

Artificial Intelligence in Classical and Quantum Photonics

Evolution towards Smart and Software-Defined Internet of Things

Intelligent nanophotonics: merging photonics and artificial intelligence at the nanoscale

Non-technological barriers: the last frontier towards AI-powered intelligent optical networks

AI/ML-as-a-Service for optical network automation: use cases and challenges

Towards explainable artificial intelligence in optical networks: the use case of lightpath QoT estimation

Towards an Intelligent Photonic System

Software-defined optical networking applications enabled by programmable integrated photonics

Artificial Intelligence in Optical Communications: From Machine Learning to Deep Learning

Machine Learning for Intelligent Optical Networks: A Comprehensive Survey

Optical networking in future-land: from optical-bypass-enabled to optical-processing-enabled paradigm