SDN Control and Modeling for Efficient Next-Generation Optical Networks

The rapid growth of global data traffic continues to challenge the scalability and effi- ciency of optical transport networks. Traditional optical systems, characterized by static configurations and high design margins, are increasingly inefficient for meeting future demands on spectrum efficiency, energy consumption, and operational flexibility. To support evolving services and traffic dynamics, optical networks require programmable and adaptive control mechanisms that can optimize the use of existing resources while maintaining interoperability across heterogeneous systems. This thesis investigates software-defined networking (SDN)–enabled approaches for enhancing the performance and flexibility of optical networks. It first examines super- channel transmission as a means to increase spectral efficiency through automated op- timization of subcarrier spacing and filter bandwidth, enabling low-margin operation without compromising the quality of transmission (QoT). The impact of margin reduction on power consumption is also investigated, revealing a trade-off between spectral and energy efficiency and highlighting the potential for power-aware super-channel opera- tion. Next, multiband (MB) transmission is explored as a strategy for capacity expansion, where thulium-doped fiber amplifiers (TDFAs) operating in the S-band are modeled and controlled via SDN using machine learning techniques to achieve dynamic and accu- rate configuration. Finally, the thesis investigates point-to-multipoint (P2MP) coherent transceivers and digital subcarrier multiplexing (DSCM) for dynamic optical access and aggregation, employing OpenConfig extensions to enable automated, vendor-neutral con- trol of network resources. In addition, fully dynamic multipoint-to-multipoint (MP2MP) operation is demonstrated using emulated DSCM transceivers. The results demonstrate the potential of SDN-based control to improve spectrum utilization, operational efficiency, and adaptability in next-generation optical networks. These studies contribute to optical network operation that is in line with the evolving re- quirements for next-generation networks, supporting more efficient, flexible, and adaptive use of network resources.