Lena Wosinska

Reliable forecasting of optical performance monitoring (OPM) parameters under non-stationary conditions is a key enabler of proactive optical network management. In this letter, we propose a forecasting enhancement framework that integrates change point and collective anomaly detection scores as auxiliary features into a sequence-to-sequence QoT forecaster. Results on production OPM data show consistent improvements in forecasting accuracy over baseline models that do not integrate the auxiliary features. As a result, the proposed framework provides a consistently positive reaction window prior to QoT violations, with observed durations ranging from 1 to 10 days across failures, enabling proactive network operation without introducing additional control loop complexity.

Fast and accurate detection of various physical layer threats that target optical networks is key to secure and reliable global communications. Conventional monitoring methods often fail to detect subtle anomalies, which requires advanced sensing techniques. Machine Learning (ML) analysis of the State of Polarization (SOP) of unmodulated signals was shown to successfully detect such disturbances. However, real-world networks typically operate with high-speed modulated signals, which may alter SOP behavior and challenge the applicability of ML techniques developed for unmodulated signals. This paper investigates the implications of signal modulation supporting high-data rates on the interpretability of SOP signatures. We perform the first experimental comparison of anomaly detection approaches based on SOP for modulated and unmodulated optical signals subjected to identical physical perturbations caused by fiber tapping and vibrations. We analyze four representative events under both signal modalities and assess the impact of modulation on SOP dynamics using a 77 km fiber link in a real-world metro network. We design four datasets that isolate, merge, and jointly classify the different signal modalities, and compare the performance of ten best-performing supervised ML techniques in each case. Our findings indicate that modulated signals tend to exhibit smoother SOP trajectories, likely due to the temporal averaging effects introduced by high symbol rates, wherein rapid symbol transitions suppress highfrequency polarization noise. Importantly, this smoothening does not obscure the slower, event-induced polarization drifts observed during physical disturbances, allowing ML models to reliably differentiate between different physical events (e.g., bending, vibrations) and signal modalities (modulated vs. unmodulated), achieving accuracy values between 97.12

Fast and accurate estimation of lightpaths’ quality of transmission (QoT) is crucial for ensuring quality of service (QoS) and seamless operation in real-world optical networks. Machine learning (ML) algorithms are promising tools for QoT estimation of lightpaths before their establishment. In multi-domain optical networks, where learned QoT estimation models must be transferred between heterogeneous environments with limited target data, deep neural networks (DNNs) have shown promising results. However, DNN-based transfer learning (TL) approaches using fine-tuned artificial neural networks (ANNs) and convolutional neural networks (CNNs), offer limited interpretability. Consequently, little insight into the decision-making process is provided, and large labeled datasets as well as high computational resources are required, limiting their suitability for real-time, large-scale deployment in production networks. To address these challenges, we propose a novel lightweight and interpretable TL framework that integrates the Boruta-SHAP algorithm for automated feature selection (FS) together with two domain adaptation (DA) techniques: Feature Augmentation and Correlation Alignment. In contrast to the existing approaches based on DNN, our strategy leverages interpretable and efficient ML models to enhance the adaptability across diverse datasets in real-world network environments. We show that our random forest (RF)-based models achieve better performance than the ANN-based models, without sacrificing the classification accuracy. The FS via Boruta-SHAP allows for reducing dimensionality as well as training and inference times up to 70.68%, and 41.64%, respectively. Our proposed framework outperforms DA baseline models achieving 99.35% accuracy improvement in domain shift. Moreover, it offers 86% accuracy with a 99.83% reduction in the size of the target domain.

The operational security of latency-sensitive networked applications is increasingly threatened by evolving malicious attacks that compromise operational integrity and network performance. Human-to-machine (H2M) applications, which rely on seamless bidirectional control signals and haptic feedback transmission, exemplify such latency-sensitive use cases. Existing learning-based malicious attack detection frameworks suffer from their reliance on pretrained datasets, making machine learning (ML) models within them ineffective against previously unseen attack patterns. As attack profiles dynamically evolve, static models become obsolete, necessitating adaptive mechanisms to maintain detection accuracy. In this context, concept drift adaptation will serve as a critical tool for enabling models to continuously adjust to changing traffic distributions and emerging attack patterns. However, real-world H2M applications lack access to accurately labeled malicious traffic data, making real-time adaptation of defense mechanisms infeasible. To address these challenges, we propose a concept drift adaptation-facilitated malicious attack defense (CDAD) framework. First, CDAD employs adaptive random forest (ARF) as an incremental learning approach, integrating an error-rate-based concept drift detection mechanism to dynamically identify evolving attack patterns and trigger adaptive model updates. Second, a haptic behavior classifier is introduced to classify expected human operator interactions and compare them with real-time haptic feedback from remote machines. This enables automated traffic relabeling, allowing CDAD to adapt to previously unseen attacks without relying on prelabeled datasets. The superior performance of CDAD over existing state-of-the-art methods is demonstrated across various malicious attack scenarios through extensive simulations. Results show that with CDAD, the attack success rate can be limited to 3% while maintaining an inference time below 1 ms, thereby ensuring effective and efficient malicious attack defense in latency-sensitive H2M applications.

Optical network infrastructure underpins global communication networks. It is exposed to various physical layer breaches, such as fiber cuts or eavesdropping via fiber bending, that may violate privacy or disrupt services. Analyses of State of Polarization (SOP) variations induced by external events, combined with Machine Learning (ML) techniques, can contribute to early identification of events and categorization of potential threats. However, real-world deployment of automated threat detection and mitigation faces many challenges, including the inconsistencies between controlled laboratory settings, often used for dataset collection for ML training, and real-world, noisy environments. In this paper, we study the detection of external disturbances in real-world fiber installations by analyzing the induced changes in the SOP of optical signals. We develop a suite of Deep Learning (DL) models, including One-Dimension (1D) Convolutional Neural Network (CNN) and fully-connected dense layers, for the detection of harmful events in noisy environments comprising a shorter (300 m) and a longer (21 km) fiber link installation, corresponding to Fiber to the Home (FTTH) and metro-scale optical paths, respectively. The proposed approach employs an optical analyzer to capture SOP changes resulting from mechanical or acoustic vibrations, as well as eavesdropping attempts. Upon careful tuning of the DL models' hyper-parameters, 98.57% accuracy is obtained for the shorter, and 92.26% for the longer link installation.

The radio-over-fiber (RoF) technology which was first introduced in the late eighties has evolved over time and is now considered as a possible solution for future wireless transport. The exponential growth in mobile user applications and their diverse requirements demand the future wireless and its transport network to be more intelligent, software-defined, and ubiquitous to provide immersive, high bandwidth, ultra-low latency, and hyper reliable communication. These evolutions mandate rethinking of the design and development of physical layer and upper layer of RoF technologies. This paper reviews the past and current developments of RoF technologies and summarizes the challenges that the technologies can potentially face in the future to support beyond 5G networks and their requirements.

Satellite optical networks combined with multi-beam technologies can be referred to as multi-beam satellite optical networks (MB-SONs). These networks are expected to play a crucial role in satellite internet, potentially achieving Gigabit/s inter-satellite (IS) communication in the future. However, the growing demand for the satellite-to-mobile communications bring a challenge for making use of hybrid IS and SG transmission resources, which is worthy of studying. Given the different characteristics of IS links and satellite-to-ground (SG) links in terms of time windows and transmission capacities, existing solutions can hardly provide a suitable option for the optimal utilization of transmission capacity in MB-SONs. In this paper, we focus on the joint scheduling of optical wavelengths in IS and multiple beams in SG for the services sent from one ground station to the other ground station (which are referred to as end-to-end services). In response to the above-mentioned different characteristics of IS and SG links, we find that at least two constraints need to be followed in the joint scheduling. Also, the length of common time window will not exceed the minimum time window of IS and SG links on the path. The second one is that the volume of transmission capacity of a path depends on the length of CTW and the minimum bandwidth of IS and SG links, which should meet the service requirements. Considering these two constraints related to the length of CTW and the volume of transmission capacity, the main contributions of this paper can be concluded: i) defining the joint time slot allocation (JTSA) problem for multiple beams and wavelengths with different transmission capacities, ii) proposing an integer linear programming (ILP) model with object to minimize the number of time slots occupied by services, iii) designing the common time window for time slot assignment (CTW-TSA) algorithm as an option in practical implementations. The proposed ILP and CTW-TSA algorithm are evaluated by comparing the simulation results with the scheme using. The simulation of the CTW-TSA algorithm was compared to separate resource scheduling (SRS) without the store-and-forward function. The results showed a reduction of almost 0.201 in service blocking probability and an increase in average bandwidth utilization of about 0.159 for IS links and 0.164 for SG uplinks/downlinks.

This Special Issue on Intelligent Networks at the Edge contains a collection of invitation-only extensions based on papers presented at the Future Optical Networks and Communications (FONC) Symposium in the IEEE Future Network World Forum (FNWF), 13–15 November 2023, Baltimore, Maryland, USA. We present a brief introduction followed by an overview of the topics covered in the papers.

Network operators must deal with Classes of Service (CoS), which have several quality requirements, such as latency, bandwidth/capacity, priority, etc. Besides, it is observed an increase in the volume of traffic that is offered to the transport network. This traffic can be affected by network natural disasters or human-made attacks. In this case, network operators must decide which services to restore, considering their different requirements. In this work, we present a Lightpath Selection Algorithm (LSA) that aims to select lightpaths to be restored after a resource crunch. This algorithm has a multicriteria decision approach considering CoS, Bandwidth, number of Hops, and Holding time. Moreover, service degradation is also considered for those lightpaths that can not be restored with full bandwidth. Results show that our proposed algorithm can improve network restorability and availability without penalizing low-level CoSs.