Network operators rely on the fault, configuration, accounting, performance, and security (FCAPS) model for efficient network management using traditional monitoring solutions that are often costly and proprietary. This paper introduces OpenNOP, an open-source, multi-layer, and multi-vendor network observability platform designed for fault detection, configuration tracking, and performance monitoring. OpenNOP collects and processes network metrics in a time-series database, enabling real-time visualization and AI-driven predictive analytics. Deployed in a multi-vendor optical transport testbed, it facilitates ML-based inference of network disturbances. OpenNOP uses scripted automation to control the generation of network disturbances and the collection of L1/L2/L3 metrics and then train and test ML models to infer the noise profile based on those metrics. By providing a scalable and extensible alternative to proprietary tools, OpenNOP advances network monitoring, predictive maintenance, and AI explainability.

According to the 3rd Generation Partnership Project, it is necessary to keep end-to-end latency to less than 5 ms in 5G mobile networks to achieve remote applications that replace on-site work and operations. The challenge here is that the bursty traffic (e.g., video) abruptly changes on a short timescale. Bursty traffic has an instantaneous high bitrate, which can exceed the link rate of the mobile midhaul (MMH) and thereby cause congestion, leading to an increase in latency. We propose a proactive congestion control method that performs parallel prediction of distributed unit (DU) traffic and fast switchover of the central unit (CU) and optical path in the MMH. Since the processing load of traffic analysis increases in accordance with the number of DUs, the CPU used for prediction becomes overloaded when many DUs are aggregated. This leads to concerns that the prediction cannot be completed in advance, that the timing of control is shifted, and that congestion occurs when switching is performed, thereby increasing the latency. The key idea of the proposed method is to optimize the number of CPU cores and threads for parallel prediction of the DU. To evaluate its feasibility and effectiveness, we experimentally optimized the number of CPU cores, thread allocations, and prediction periods and then estimated the maximum number of simultaneous predictions. The experimental results for a simple condition where video traffic and background traffic are mixed show that our prototype controller successfully achieved a latency at the MMH as low as 1 ms with appropriate congestion control. Our findings demonstrate the latency reduction effect of the proposed congestion control method for MMH in an aggregated DU configuration for applications with severe latency requirements such as remote control.

The rapid evolution of wireless communication technologies has led to the development of Beyond 5G (B5G) networks, aiming to deliver ultra-low latency, high-capacity connectivity, and scalable deployments for applications including autonomous systems and immersive experiences. A critical challenge lies in efficiently designing the radio access network (RAN) and its transport infrastructure. The shift to open RAN (O-RAN) architectures, with disaggregated centralized units (CUs), distributed units (DUs), and radio units (RUs), offers a promising solution. However, optimal placement of these components and transport network design remains a challenge, especially in diverse and cost-sensitive environments. This paper proposes a comprehensive optimization framework for B5G network design, jointly optimizing the placement of O-RAN components and optical transport networks using Mixed Integer Linear Programming (MILP). By integrating geographic information system (GIS) data, the framework considers real-world constraints, including terrain, infrastructure, user distribution, coverage requirements, and capacity requirements, enabling practical and cost-effective designs. The proposed optimization framework allows exploration of the optimal planning solutions under diverse configurations, which helps analyze different network deployment approaches for their scalability and efficiency. The evaluation results demonstrate the effectiveness and versatility of the framework in reducing costs under diverse deployment scenarios while meeting performance and coverage requirements, offering valuable insights for network operators and planners.

Multi-band transmission is gaining traction as a promising solution for coping with the ever-increasing capacity demands of optical networks. In multi-band networks, long-term network planning must take into account the signal-to-noise ratio (SNR) degradations caused by interchannel stimulated Raman scattering arising when multiple lightpaths provisioned in the extended spectrum bands propagate in the same fiber. To avoid these degradations, in this study, we investigate the benefits of selectively upgrading a subset of the fiber spans with an additional fiber, which allows us to separate the spectrum bands and apply dual-fiber distributed Raman amplification (DF-DRA) in a C+L+S system. The deployment of DF-DRA, combined with EDFAs for the C- and L-bands and TDFAs for the S-band, reduces the overall amplification noise figure, thereby improving the SNR of lightpaths. Additionally, the separation of spectrum bands reduces pump-to-pump and signal-to-signal power transfer, further contributing to SNR improvement. We develop a numerical solver for setting the optimal Raman pump configurations, and we propose different placement strategies for hybrid EDFA/TDFA-Raman amplification and 3R regenerators. Our extensive numerical simulations on incremental traffic show that DF-DRA-enabled multi-band networks can yield up to a complete elimination of 3R regenerators by upgrading 30% of the network spans while having up to a 35% throughput increase.

This paper proposes a 6G architecture that adopts a multi-technology transport network integrating THz links and optical network technologies to interconnect the 6G radio access network and core domains. The proposed solution operates in a self-organized manner, taking advantage of the software defined networking (SDN) control of the optical network, while suitable SDN control is purposely developed for the THz solution. An end-to-end (E2E) Service Management and Orchestration (SMO) layer is adopted, offering intelligence capabilities in service provisioning and resource allocation across the 6G infrastructure through the adoption of a federated learning (FL) scheme. The proposed architecture and the developed control scheme are validated through an experimental demonstration. To the best of the authors’ knowledge, this is the first experimental demonstration of an autonomously controlled 6G network implementation integrating optical network technologies and THz links in a common transport network. The developed intelligent management framework empowered by FL to jointly optimize THz and multi-vendor optical network equipment supporting 6G services is also experimentally showcased.

White Rabbit has emerged as a promising technology for time-scale dissemination, providing resilience to satellite techniques. To achieve state-of-the-art synchronicity at the sub-nanosecond level in a White Rabbit network, a complete characterization of physical effects, components, and the reference time scale is required in order to remove systematic offsets not compensated by the White Rabbit protocol. Due to this complexity, only a few long-range White Rabbit networks can claim synchronicity to a national time scale on a sub-nanosecond level. We present here such a network with the vocation to disseminate the Swiss official time-scale UTC(CH). To that end, the White Rabbit network layer is multiplexed into two dark channels in the L-band of a deployed telecommunication fiber network covering nearly 500 km of fiber. We characterize its performance in terms of stability and accuracy and find an uncertainty of 289–415 ps. This validates its potential to disseminate UTC(CH) at the sub-nanosecond accuracy and precision level. The results confirm that White Rabbit provides the necessary performance to distribute and compare the national time-scale UTC(CH) over optical fiber networks at the national and eventually at international levels.

Fiber-optic communication systems serve as the backbone of modern data communication networks, with increasing demands on their reliability and robustness in various emerging applications. A key challenge in ensuring reliable fiber-optic transmission lies in addressing fiber anomalies, which can cause signal degradation, service disruptions, and even system failures. However, the current anomaly detection method is too complex to be implemented in deployed networks or consumes too much time during detection. This paper proposes a simple and effective fiber anomaly detection method for C+L-band fiber-optic communication systems, leveraging the spectral tilt induced by the stimulated Raman scattering (SRS) effect. The method reconstructs the spectral tilt along an anomalous fiber link by analyzing the input and output power profile, easily obtainable from optical channel monitors (OCMs), enabling anomaly localization and loss quantification through forward and backward spectral tilt comparison. The performance and robustness of the proposed method are studied and discussed in actual scenarios such as erbium-doped fiber amplifier gain fluctuations, OCM measurement errors, parameter inaccuracies, and modeling inaccuracies. In addition to being able to detect a single anomaly, the process of using the proposed method to achieve multiple anomaly detection links is also discussed according to the chronological order of anomalies. Experimental validation on a multi-span C+L-band transmission system achieves a maximum localization error of 2.06 km and a maximum loss estimation error of 0.14 dB within 2 s. This work underscores the potential of exploiting inherent physical phenomena like SRS for fast anomaly detection in deployed optical networks.

Multi-chip graphics processing units (GPUs) interconnected by a photonic network-on-wafer are a promising technology to further increase the performance of GPUs. The network control algorithm managing dynamic bandwidth allocation (DBA) in this network needs to execute very frequently so that resources can be optimally used. This algorithm relies on edge coloring bipartite multigraphs to translate inter-chip bandwidth demands into updated routing tables for the GPU chips and optical switches in the network. In this work, we design fast edge coloring algorithms, both approximate and exact, for bipartite multigraphs. These algorithms are tailored to the high edge multiplicities of the multigraphs in this research. The runtimes are optimized by using efficient data structures and introducing pre- and post-processing. These new algorithms are up to 20× faster than the state-of-the-art baseline algorithm. New simulations show that, with such low reconfiguration periods, DBA has the potential to double the performance of a high-traffic GPU workload compared to a static network with the same bandwidth.

The advancement of machine learning (ML)-assisted solutions for monitoring and performance analysis in space-division multiplexing (SDM) networks has been significantly constrained by a shortage of large, well-structured, and publicly available datasets. As a result, researchers often rely on custom-built datasets, making reproducibility difficult and complicating cross-comparisons of proposed methods. To address this gap, we introduce 18 novel, to our knowledge, and publicly available quality of transmission (QoT) datasets, designed specifically for SDM networks. These datasets cover a wide range of SDM configurations, incorporating multiple fiber types, switching strategies, and two distinct network topologies. By offering consistent benchmarks, these datasets aim to support the development of ML-driven automation in SDM networks, making studies more efficient, reliable, and comparable. To demonstrate practical applications, we developed and evaluated two ML models—a classification model and a regression model—using these datasets to predict QoT outcomes. The findings illustrate the importance of the proposed datasets in benchmarking various ML-based approaches, facilitating comparison and validation across studies, and advancing ML-based network automation. This work accelerates progress toward more operationally efficient and high-performing SDM networks.

With widespread 10 Gbps-capable PON technologies in the access network and Wi-Fi 6E/7 in the home network, final customers can easily enjoy a gigabit experience at home. However, with such enormous bandwidth provided, customers are not always satisfied with their experience, particularly for real-time services, such as VR gaming, cloud gaming, or visio-conferencing, due to network latency issues. In this paper, we focus on optimizing congestion latency that can arise in both home gateway and access networks while keeping an eye on bandwidth-sharing fairness without strictly prioritizing certain traffic. To assess end-to-end performance, our setup includes Internet servers, access/home networks, and customers’ equipment. Transport layer protocols in the TCP/IP model are deployed in both Internet servers and end-user equipment. TCP Prague is implemented to represent time-sensitive services that require a high bitrate, such as VR, and TCP Cubic is used for other common services without low-latency criteria. We create a two-level congestion scenario, configuring different QoS parameters within access equipment and home gateways to compare performance. In our best case, the latency of time-sensitive traffic is at least 10 times lower than other applications and gains more than 30 ms E2E RTT during full congestion scenarios.