This special issue presents the extended versions of selected papers presented at the 27th International Conference on Optical Network Design and Modeling (ONDM 2023), which took place on 8–11 May 2023 at the University of Coimbra, Coimbra, Portugal. The conference delved into pioneering research within well-established realms of optical networking, enabling a diverse range of novel services and applications. The presented research included cutting-edge topics like 5G and beyond, data center networking, the Internet of Things (IoT), cloud and edge computing, content delivery, big data analytics, network telemetry, real-time monitoring, autonomic networking, artificial intelligence, machine learning-driven networks, visible light networks, and quantum secured networks. The program featured 34 contributed papers and 6 posters, as well as 3 keynotes, 12 invited talks, 3 tutorials, and 2 workshops. With over 90 participants from all over the world, ONDM 2023 was a very successful conference. The papers in this special issue cover some of the current key trends in optical networking research, namely quality of transmission (QoT) estimation, power awareness, network resilience, and quantum key distribution network design. QoT works address detection, modeling, and mitigation of QoT fluctuations [1], analysis of the importance of QoT estimation to perform optical line control [2], and optimization of transport network planning [3]. A power-aware super-channel optimization procedure is proposed [4] in an elastic optical network (EON) controlled by software-defined networking (SDN). Improving network resilience is tackled by link availability modeling [5], failure recovery in IP over wavelength division multiplexing (IPoWDM) [6], and methods for disaster recovery in multi-carrier networks [7]. Finally, strategies to upgrade long-haul quantum key distribution networks, are discussed [8].

The monitoring data from deployed backbone networks indicates that many lightpaths experience significant fluctuations in their QoT, exceeding a range of 1 dB. In [1], the authors investigate the hypothesis that these fluctuations are caused by anomalies in polarization dependent loss (PDL), which is equivalent to a single large PDL element. A method based on closed-loop automation is proposed and experimentally tested to stabilize the QoT at a maximal level for lightpaths impaired by such a PDL anomaly.

The increase of bandwidth requirements encourages optimization of the spectrum utilization, which needs accurate calibration of the QoT. In [2], the authors evaluate the QoT estimation of three different models for two different types of monitoring equipment (power meters and emulated optical channel monitors) for an EDFA-based optical line system. The study considers six different optimization methods and evaluates not only the accuracy but also the complexity and computation time as well as the most influential monitoring parameters.

Flexible grid EONs have recently been widely deployed to support the growing demand for bandwidth-intensive applications. Optimal utilization of EONs is required to delay costly network upgrades and reduce cost and power consumption. In [3], the authors investigate the impact of increased configuration granularity on optical networks. They consider the practical implementation of bandwidth-variable transceiver (BVT) configurations for estimating the required signal-to-noise ratio. In addition, an optimization algorithm is presented that selects the most efficient configuration for each considered data rate and bandwidth combination.

Telecommunication networks are major consumers of power. This highlights the importance of developing more energy-efficient technologies and network management strategies. A power-aware super-channel optimization procedure to identify the most power- and spectrum-efficient super-channel configurations was experimentally demonstrated in [4]. Furthermore, experiments on the power consumption of a thulium doped fiber amplifier (TDFA) allowed the conclusion that its configuration could be optimized through machine learning techniques.

Today’s society depends on continued telecommunications services. The availability of optical links is one of the key factors for dependable services. However, modeling link availability remains a difficult problem. A model to estimate optical network link availability, combining data and a hierarchical Bayesian model, is presented in [5]. Results show the proposed model has higher accuracy than other baseline approaches such as empirical interval availability or polynomial and reciprocal regressions.

The work [6] analyses failure recovery in the so-called MANTRA (metaverse ready architectures for open transport) architecture, which relies on IPoWDM nodes and 400ZR/ZR+ pluggables. Three procedures to coordinate and control these multi-layer IPoWDM nodes were designed and experimentally validated on an experimental testbed comprising a sliced IPoWDM node running an extended version of the open-source SONiC network operating system. The experiment showed that schemes where the IPoWDM nodes with alternative paths proved to be the best performing solution, with very relevant restoration time reductions with respect to the only-optical restoration mechanism.

Large-scale carrier network resilience is a primary societal concern. In multi-carrier network ecosystems, cooperation among different carriers is crucial to achieving resilience against large-scale failures. However, cooperation between carriers can be challenging when carriers hesitate to disclose confidential information, such as detailed resource availability. The study in [7] investigates how to perform carrier cooperative recovery in the event of large-scale failures or disasters. The authors propose a two-stage carrier-carrier cooperative recovery plan that incorporates coordinated scheduling for faster recovery.

The advancements in quantum computing forced the development of quantum key distribution networks (QKDNs), for preserving privacy and security in optical transport networks (OTNs). In [8], an algorithm for topology optimization was proposed, and upgrade strategies for QKDNs were evaluated. It was concluded that placing trusted nodes (TNs) was the most effective strategy, albeit environmental circumstances may constrain it.

We expect the papers presented will encourage JOCN readers to embark on innovative and pioneering research initiatives.