Journal of Optical Communications and Networking

Machine learning (ML)-based quality-of-transmission (QoT) estimation tools will be desirable for operating virtual network topologies (VNTs) that disclose only abstracted views of connectivity and resource availability to tenants. Conventional ML-based solutions rely on laborious human effort on model selection, parameter tuning, and so forth, which can cause prolonged model building time. This paper exploits the learning-to-learn nature by meta learning to pursue automated provisioning of QoT estimators for a dynamic VNT configuration in optical networks. In particular, we first propose a graph neural network (GNN) design for network-wide QoT estimation. The proposed design learns global VNT representations by disseminating and merging features of virtual nodes (conveying transmitter-side configurations) and links (characterizing physical line systems) according to the routing schemes used. Consequently, the GNN is able to predict the QoT for all the end-to-end connections in a VNT concurrently. A distributed collaborative learning method is also applied for preserving data confidentiality. We train a meta GNN with meta learning to acquire knowledge generalizable across tasks and realize automated QoT estimator provisioning by fine tuning the meta model with a few new samples for each incoming VNT request. Simulation results using data from two realistic topologies show our proposal can generalize QoT estimation for VNTs of arbitrary structures and improves the estimation accuracy by up to 18.7% when compared with the baseline.

Fiber-based quantum key distribution (QKD) systems are mature and commercialized, but their integration into existing optical networks is crucial for their widespread use, in particular in passive optical networks (PONs) if end-to-end quantum-secured communications are to be addressed. While discrete-variable QKD coexistence with classical channels is well-studied in point-to-point links, its performance in point-to-multipoint topologies like PONs has received less attention. We thus developed a numerical tool to estimate quantum-available bandwidth and maximum link lengths for QKD systems in single-fiber PON architectures in coexistence with GPON, XG-PON, NG-PON2, and HS-PON standards. The QKD channel performance is obtained by setting thresholds on the quantum bit error rate and the secret key rate, ultimately limited by spontaneous Raman scattering noise and high optical distribution network losses. We perform a comparison between the performance obtained assuming the asymptotic infinite-key generation rate or taking into account actual implementations in the finite-key regime. We evidence that proper design rules can be obtained as a function of both classical and quantum system parameters to support end-to-end quantum security services in existing optical networks.

Topology Bench is a comprehensive topology dataset designed to accelerate benchmarking studies in optical networks. The dataset, focusing on core optical networks, comprises publicly accessible and ready-to-use topologies, including (a) 105 georeferenced real-world optical networks and (b) 270,900 validated synthetic topologies. Prior research on real-world core optical networks has been characterized by fragmented open data sources and disparate individual studies. Moreover, previous efforts have notably failed to provide synthetic data at a scale comparable to our present study. Topology Bench addresses this limitation, offering a unified resource, and represents a 61.5% increase in spatially referenced real-world optical networks. To benchmark and identify the fundamental nature of optical network topologies through the lens of graph-theoretical analysis, we analyze both real and synthetic networks using structural, spatial, and spectral metrics. Our comparative analysis identifies constraints in real optical network diversity and illustrates how synthetic networks can complement and expand the range of topologies available for use. Currently, topologies are selected based on subjective criteria, such as preference, data availability, or perceived suitability, leading to potential biases and limited representativeness. Our framework enhances the generalizability of optical network research by providing a more objective and systematic approach to topology selection. A statistical and correlation analysis reveals the quantitative range of all of these graph metrics and the relationships between them. Finally, we apply unsupervised machine learning to cluster real-world topologies into distinctive groups based on nine optimal graph metrics using K-means. It employs a two-step optimization process: optimal features are selected by maximizing feature uniqueness through principal component analysis, and the optimal number of clusters is determined by maximizing decision boundary distances via support vector machines. We conclude the analysis by providing guidance on how to use such clusters to select a diverse set of topologies for future studies. Topology Bench, openly available via Dataset 1 (https://zenodo.org/records/13921775) and Code 1 (https://github.com/TopologyBench), promotes accessibility, consistency, and reproducibility.

We explore optical switching to extend network programmability to the physical layer and discuss applications of a Layer-1 software-defined network (SDN) in AI/HPC clusters. In this context we identify two applications for optical circuit switches (OCSs): failure resilience and reconfigurable topologies for deep learning workloads. We present experimental results from a DGX-based testbed towards improving failure resilience and a simulation analysis for efficient deep learning training in AI clusters.

This paper presents a comprehensive overview of the emerging coherent passive optical network (CPON) technology and its role in the evolution of next-generation PON architectures. After reviewing the fundamental principles of traditional time-division multiplexed PON and tracking its development across successive standards, the unique benefits of applying coherent detection to PONs are examined. These include enhanced reach, increased split ratios, and improved overall network scalability. The paper explores various use cases, deployment scenarios, and architectural options for CPONs. Critical enabling technologies driving CPON development are analyzed, including upstream preamble design and burst-mode detection, techniques for transceiver cost reduction and implementation simplification, forward error correction, and out-of-band communication channels. Finally, the major industry standardization efforts currently underway to specify CPON across the physical, media access control, and system architecture layers are reviewed. By cohesively covering CPON’s evolution from core concepts to real-world specifications, this tutorial paper provides a definitive reference on this disruptive access network technology.

Effective management of end-to-end 6G network services is crucial, with peak capacity requirements for 6G transport connections expected to exceed 1 Tb/s. As demand for high bandwidth rises, there is a growing necessity for high-capacity optical fiber links, including ultra-wideband (UWB) and multiple fiber links within the network. Scaling up to accommodate these demands, designing wavelength-selective switches (WSSs) for such networks significantly increases the port count. To tackle this issue, we propose various multi-granular optical node (MG-ON) architectures utilizing heterogeneous wavelength, waveband, and spatial switching. We evaluate these architectures’ performance against high-capacity wavelength division multiplexed (WDM) networks through various simulation parameters.

Digital subcarrier multiplexing (DSCM) based coherent point-to-multipoint transceivers (P2MP-TRXs) are promising for addressing the shift in traffic patterns from point-to-point (P2P) to hub-and-spoke (H&S), and their application in wavelength-switched optical networks (WSONs) can potentially offer enhanced flexibility and efficiency in handling the mixed traffic therein. In this paper, we study how to secure the survivability of P2MP-TRX-based WSONs against packet layer failures with cross-layer restoration (CLR). By analyzing the unique features of P2MP-TRXs, we first design three CLR strategies to restore the traffic affected by packet layer failure(s) and then formulate an integer linear programming (ILP) model to leverage them for cost-effective CLR, i.e., minimizing the cost introduced during the CLR process. Next, we propose a time-efficient heuristic, namely, hHAG-DP, which leverages hybrid dynamic programming (DP) and a hierarchical auxiliary graph (HAG) to find cost-effective CLR schemes quickly. Extensive simulations confirm the effectiveness of our proposals.

The use of optical access networks with aerial-deployed fiber for deriving maps of environmental temperature is investigated. Telecom operators have thousands of kilometers of deployed fiber to provide last-mile broadband services, which could be leveraged to extract temperature information with no additional cost since data are already available as part of the physical layer operations, administration, and maintenance (PLOAM) traffic. Here, it is shown how this information can be used to develop maps of environmental temperature as a method to complement present weather observation platforms. Preliminary experimental results with a G.984 passive optical network (PON) in operation show the feasibility of the technique.

We propose a pointer network-based joint routing, modulation format, and spectrum allocation (PtrNet-RMSA) scheme to optimize the network capacity in elastic optical networks (EONs). Specifically, for a service request, EON spectra are first split into multiple spectrum window planes (SWPs), regarding various numbers of contained frequency slots (FSs). Then, over each SWP, a lightpath with an optimal estimated quality of transmission (QoT) is generated using a PtrNet-based routing agent. Further, different modulation formats are verified for the generated lightpath over the considered SWP, with respect to a calculated QoT threshold, to check whether a specific preforward error correction (FEC) requirement can be satisfied. Finally, two heuristic approaches, first fit and highest fit, are applied in the joint RMSA to select an optimal lightpath of an SWP and modulation format. To evaluate the efficiency of the PtrNet-RMSA scheme, extensive simulations were conducted in EONs with dynamic traffic. Results show that the proposed PtrNet-RMSA scheme can significantly improve the network capacity with various network topologies, compared with two benchmarks.

Quality of transmission (QoT) prediction is a fundamental function in optical networks. It is typically embedded within a digital twin and used for operational tasks, including service establishment, service rerouting, and (per-channel or per-amplifier) power management to optimize the working point of services and hence to maximize their capacity. Inaccuracy in QoT prediction results in additional, unwanted design margins. A key contributor to QoT inaccuracy is the uncertain knowledge of fiber insertion loss, e.g., the attenuation due to connector losses at the beginning or at the end of each fiber span, as such loss cannot be directly monitored. Indeed, insertion losses drive the choice of the launch power in fiber spans, which in turn drive key physical effects, including the Kerr and stimulated Raman scattering (SRS) effects, which affect services’ QoT. It is thus important to estimate (and detect possibly anomalous) fiber insertion losses at each span. We thereby propose a novel active input refinement (AIR) technique using active probing to estimate insertion losses in C and C + L systems. Here, active probing consists of adjusting amplifier gains span by span to slightly alter SRS. The amount of adjustment must be sufficient to be measurable (such that insertion losses can be inferred from the measures) but small enough to have a negligible impact on running services in a live network. The method is validated by simulations on a European network with 30 optical multiplex sections (OMSs) in C and C + L configurations and by lab experiments on a C-band network, demonstrating that AIR significantly improves insertion loss estimation, network QoT optimization, and QoT prediction compared with other state-of-the-art monitoring techniques. This work underscores the critical role of accurate estimation of QoT inputs in enhancing optical network performance.

With the ability to provide worldwide communication coverage, satellite networks are drawing greater attention. The translucent optical payload enables the implementation of IP-over-WDM satellite optical networks (SONs), which can achieve great bandwidth capacity while providing the flexibility of IP routing. The rechargeable battery is the sole energy support for satellites in the eclipse region. Unrestrained discharge will accelerate battery aging and shorten the satellite operation period, causing extremely high expenditure costs. Satellite movement causes time-scheduled energy supply and traffic fluctuation, complicating the problem of energy consumption in IP-over-WDM SONs. This paper studies green traffic grooming (GTG) in IP-over-WDM SONs from the perspective of battery lifetime consumption (BLC). A grooming graph is designed to implement GTG with the physical impairment constraint in IP-over-WDM SONs, and battery-aware GTG (BA-GTG) and time-aware GTG (TA-GTG) are proposed by taking battery information and time information as prior knowledge. Numerical results indicate that BA-GTG and TA-GTG, especially the latter, can effectively reduce BLC. In addition, multiple link configurations are set in performance comparison to evaluate the effect of the physical impairment on battery efficiency in IP-over-WDM SONs.

Emerging use cases with demanding bandwidth and latency requirements, as well as the challenge of reducing power consumption, are driving the need for evolution in optical network architectures. An all-photonics metro-access converged network (APN) aims to actualize a flat architecture by expanding dense wavelength-division-multiplexing (DWDM) metro networks into access areas. The APN flexibly and dynamically provides optical connectivity between any two points, even across the boundaries between access and metro areas according to individual application requirements and traffic-load status. To actualize and further evolve the APN concept, several technical challenges regarding access nodes, defined as Photonic Gateways (GWs), still remain. From an access node functionality perspective, first, the Photonic GW should forward various types of optical paths. Unlike reconfigurable optical add/drop multiplexers in current metro networks, which are specifically designed to cross-connect DWDM signals, the Photonic GW needs to handle various lights and optical signals, including short-reach applications and emerging non-communication use cases. Second, the Photonic GW should provide remote control channels to user terminals (UTs) in a more scalable and cost-effective manner within the node-footprint and power-consumption constraints. Remote and in-channel UT control is required to place flexibly the endpoints of a wavelength path, i.e., UT, beyond the control-plane network. Then, from the controller perspective, the physical connectivity between the newly connected UT and the access-side port of the Photonic GW should be autonomously identified for plug-and-play operation. Since UTs are widely distributed within an access area, there is a need for an initial connection method that does not require timing adjustments to connect to the APN between newly connected UTs. This paper presents an extension to the APN architecture that allows the Photonic GW to increase the types of accommodable optical paths and to enhance the scale of remote UT control. This paper also proposes an advanced initial connection method that works even when multiple UTs are simultaneously connected to the APN. The extension to the APN architecture and the initial connection method are verified through experiments based on a Photonic GW prototype that fully complies with the extended APN architecture and comprises four functionally disaggregated units, 100-Gb/s C-band DWDM digital coherent UTs, and 25-Gb/s O-band non-DWDM intensity modulation and direct detection UTs. A 10-gigabit-capable symmetric passive optical network is adopted for remote UT control. The proposed initial connection method eliminates the connection interval of 6 s or more between newly connected UTs and achieves guard time-free operation.

In this work, we experimentally demonstrate a ${100}\;{\rm Gb/s/}\lambda$ downstream transmission link for coherent passive optical networks (PONs) up to 50 km, achieving an optical power budget of 29 dB through polarization multiplexing (PolMux) of two 50 Gb/s channels using multiband carrierless amplitude phase modulation (multiCAP) and optical single side band (OSSB) modulation. Additionally, we introduce a separate PolMux 50 Gb/s link that presents an optical power budget of 38.7 dB. Both links have been achieved using a simplified polarization-demultiplexing heterodyne coherent receiver. The robustness of the system is experimentally evaluated by analyzing its response to various input states of polarization. The transmission has been accomplished using 10 GHz electrical bandwidth devices at both the transmission and receiving ends, thereby paving the way for low-cost 100G links suitable for applications such as PONs.

The first part of this special issue contains five papers addressing different aspects of good benchmarking practices in optical network research. We start with a brief introduction, followed by an overview of the papers and an invitation to contribute to the second part.