We evaluate the advantages of using the extra dimension introduced by space-division multiplexing (SDM) for dynamic bandwidth-allocation purposes in a flexible optical network. In that respect, we aim to compare spectral and spatial super-channel (Sp-Ch) allocation policies in an SDM network based on bundles of SMFs (to eliminate coupling between spatial dimensions from the study) and to investigate the role of modulation format selection in the blocking probability performance with an emphasis on the spectral efficiency (SE)/reach tradeoff for different multiline-rate scenarios, created either by changing the number of sub-channels (Sb-Ch), or by employing different modulation formats. Our network-performance results show that DP-8QAM -in a multichannel (MC) single-modulation-format system assuming ITU-T 50-GHz WDM Sb-Ch spectrum occupation-offers the best compromise between SE and optical reach for both spectral and spatial Sp-Ch allocation policies. They also reveal that an MC multimodulation-format system improves the network performance, particularly for spectral Sp-Ch allocation with Sb-Ch spectrum occupation of 37.5 GHz on the 12.5-GHz grid. Additionally, as another important contribution of the paper, we investigate, for spatial Sp-Ch allocation, the performance of several SDM switching options: independent switching (InS), which offers highest flexibility, joint-switching (JoS), which routes all spatial modes as a single entity, and fractional-joint switching, which separates out the spatial modes into sub-sets of spatial modes which are routed independently. JoS is proved to offer a similar performance to that of InS for particular network load profiles, while allowing a significant reduction in the number of wavelength-selective switches.

The availability of coarse-resolution cost-effective Optical Spectrum Analyzers (OSA) allows its widespread deployment in operators’ networks. In this paper, we explore several machine learning approaches for soft-failure detection, identification and localization that take advantage of OSAs. In particular, we present three different solutions for the two most common filter-related soft-failures; filter shift and tight filtering which noticeably deform the expected shape of the optical spectrum. However, filter cascading is a key challenge as it affects the shape of the optical spectrum similarly to tight filtering; the approaches are specifically designed to avoid the misclassification of properly operating signals when normal filter cascading effects are present. The proposed solutions are: $i) multi-classifier$ approach, which uses features extracted directly from the optical spectrum, $ii) single-classifier$ approach, which uses pre-processed features to compensate for filter cascading, and $iii) residual-based$ approach, which uses a residual signal computed from subtracting the acquired single by OSAs from an expected signal synthetically generated. Extensive numerical results are ultimately presented to compare the performance of the proposed approaches in terms of accuracy and robustness.

In elastic optical networks (EONs), effective soft failure localization is of paramount importance to early detection of service level agreement violations while anticipating possible hard failure events. So far, failure localization techniques have been proposed and deployed mainly for hard failures, while significant work is still required to provide effective and automated solutions for soft failures, both during commissioning testing and in-operation phases. In this paper, we focus on soft failure localization in EONs by proposing two techniques for active monitoring during commissioning testing and for passive in-operation monitoring. The techniques rely on specifically designed low-cost optical testing channel (OTC) modules and on the widespread deployment of cost-effective optical spectrum analyzers (OSAs). The retrieved optical parameters are elaborated by machine learning-based algorithms running in the agent’s node and in the network controller. In particular, the Testing optIcal Switching at connection SetUp timE (TISSUE) algorithm is proposed to localize soft failures by elaborating the estimated bit-error rate (BER) values provided by the OTC module. In addition, the FailurE causE Localization for optIcal NetworkinG (FEELING) algorithm is proposed to localize failures affecting a lightpath using OSAs. Extensive simulation results are presented, showing the effectiveness of the TISSUE algorithm in properly exploiting OTC information to assess BER performance of quadrature-phase-shift-keying-modulated signals, and the high accuracy of the FEELING algorithm to correctly detect soft failures as laser drift, filter shift, and tight filtering.

A successful migration from current C-band based optical networks to a multiband scenario primarily depends on the development of solutions that can reliably measure physical properties of optical links over broad spectral transmission windows. Additionally, these solutions must be capable of delivering wavelength-dependent and spatially-resolved indicators that can empower network operators to identify faults before they lead to severe service disruptions. Recently, the exploitation of receiver based digital signal processing as a tool for optical performance monitoring has gained tremendous popularity. One successful example is the so-called in-situ power profile estimator, which can reconstruct the per-channel longitudinal power profile along the optical fiber link solely processing the received signal samples. In this work, we propose a novel application for the in-situ power profile estimator by harnessing it on multiple wavelengths to accurately estimate the spectral gain profile of C+L-band in-line Erbium-doped fiber amplifiers deployed in a 280-km single mode fiber link. Furthermore, we show how this scheme can be efficiently used to detect amplification-related anomalies, such as gain tilt and narrowband gain compression. In our measurements, we achieved a sub-dB estimation accuracy by comparing the proposed gain extraction approach with the back-to-back characterization obtained from an optical spectrum analyzer.

Spatially integrated switching architectures have been recently investigated in an attempt to provide switching capability for networks based on spatial division multiplexing (SDM) fibers, as well as to reduce the implementation cost. These architectures rely on the following switching paradigms, furnishing different degrees of spectral and spatial switching granularity: independent switching, which offers full spatial-spectral flexibility; joint-switching, which treats all spatial modes as a single entity; and fractional-joint switching, whereby subgroups of spatial modes are switched together as independent units. The last two paradigms are categorized as spatial group switching solutions since the spatial resources (modes, cores, or single-mode fibers) are switched in groups. In this paper, we compare the performance (in terms of spectral utilization, data occupancy, and network switching infrastructure cost) of the SDM switching paradigms listed above for varying spatial and spectral switching granularities in a network planning scenario. The spatial granularity is related to the grouping of the spatial resources, whereas the spectral granularity depends on the channel baud rate and the spectral resolution supported by wavelength selective switches (WSS). We consider two WSS technologies for handling of the SDM switching paradigms: 1) the current WSS realization, 2) WSS technology with a factor-two resolution improvement. Bundles of single-mode fibers are assumed across all links as a near-term SDM solution. Results show that the performance of all switching paradigms converge as the size of the traffic demands increases, but finer spatial and spectral granularity can lead to significant performance improvement for small traffic demands. Additionally, we demonstrate that spectral switching granularity must be adaptable with respect to the size of the traffic in order to have a globally optimum spectrum utilization in an SDM network. Finally, we calculate the number of required WSSs and their port count for each of the switching architectures under evaluation, and estimate the switching-related cost of an SDM network, assuming the current WSS realization as well as the improved resolution WSS technology.

Autonomic optical transmission and networking requires machine learning (ML) models to be trained with large datasets. However, the availability of enough real data to produce accurate ML models is rarely ensured since new optical equipment and techniques are continuously being deployed in the network. One option is to generate data from simulations and lab experiments, but such data could not cover the whole features space and would translate into inaccuracies in the ML models. In this paper, we propose an ML-based algorithm life cycle to facilitate ML deployment in real operator networks. The dataset for ML training can be initially populated based on the results from simulations and lab experiments. Once ML models are generated, ML retraining can be performed after inaccuracies are detected to improve their precision. Illustrative numerical results show the benefits of the proposed learning cycle for general use cases. In addition, two specific use cases are proposed and demonstrated that implement different learning strategies: (i) a two-phase strategy performing out-of-field training using data from simulations and lab experiments with generic equipment, followed by an in-field adaptation to support heterogeneous equipment (the accuracy of this strategy is shown for a use case of failure detection and identification), and (ii) in-field retraining, where ML models are retrained after detecting model inaccuracies. Different approaches are analyzed and evaluated for a use case of autonomic transmission, where results show the significant benefits of collective learning.