Lightpath Quality Of Transmission Research Articles

Towards explainable artificial intelligence in optical networks: the use case of lightpath QoT estimation

Artificial intelligence (AI) and machine learning (ML) continue to demonstrate substantial capabilities in solving a wide range of optical-network-related tasks such as fault management, resource allocation, and lightpath quality of transmission (QoT) estimation. However, the focus of the research community has been centered on ML models’ predictive capabilities, neglecting aspects related to models’ understanding, i.e., to interpret how the model reasons and makes its predictions. This lack of transparency hinders the understanding of a model’s behavior and prevents operators from judging, and hence trusting, the model’s decisions. To mitigate the lack of transparency and trust in ML, explainable AI (XAI) frameworks can be leveraged to explain how a model correlates input features to its outputs. In this paper, we focus on the application of XAI to lightpath QoT estimation. In particular, we exploit Shapley additive explanations (SHAP) as the XAI framework. Before presenting our analysis, we provide a brief overview of XAI and SHAP, then discuss the benefits of the application of XAI in networking and survey studies that apply XAI to networking tasks. Then, we model the lightpath QoT estimation problem as a supervised binary classification task to predict whether the value of the bit error rate associated with a lightpath is below or above a reference acceptability threshold and train an ML extreme gradient boosting model as the classifier. Finally, we demonstrate how to apply SHAP to extract insights about the model and to inspect misclassifications.

Survey on the Use of Machine Learning for Quality of Transmission Estimation in Optical Transport Networks

Estimating the Quality of Transmission (QoT) of the optical signal from source to destination nodes is the cornerstone of design engineering and service provisioning in optical transport networks. Recent studies have turned to Machine Learning (ML) techniques to improve the accuracy of QoT estimation. In this paper, we survey the literature on this topic and classify the studies into categories based on their scope. Accordingly, we distinguish four categories of ML-based solutions: i) check lightpath feasibility, ii) estimate a lightpath's QoT, iii) enhance existing analytical models and iv) improve model generalization. We describe the proposed solutions in each category in terms of ML algorithms, inputs/outputs of the models, source of data and performance evaluation. Deploying a ML-based solution in the real field is not straightforward and presents several challenges. Therefore, we also discuss from an operator's perspective the potential of these solutions for real-field deployment.

Machine-Learning-Based Lightpath QoT Estimation and Forecasting

Machine learning (ML) is more and more used to address the challenges of managing the physical layer of increasingly heterogeneous and complex optical networks. In this tutorial, we illustrate how simple and more sophisticated machine learning methods can be used in lightpath quality of transmission (QoT) estimation and forecast tasks. We also discuss data processing strategies with the aim to determine relevant features to feed the ML classifiers and predictors. We then introduce a preliminary study on the application of transfer learning to try to overcome the scarcity of field data.

Dynamic impairment-aware RMCSA in multi-core fiber-based elastic optical networks

Space division multiplex elastic optical networks (SDM-EONs) based on multi-core fiber (MCF) have been regarded as one of the promising solutions to address the future network capacity crunch. Inter-core crosstalk (XT) and intra-core physical layer impairments (PLI) exist in the MCF and degrade the quality of transmission (QoT) for lightpaths. Meanwhile, the establishment of new lightpaths affects the QoT of existing lightpaths due to dynamic spectrum occupancy in the network. This paper focuses on resource allocation under the constraint of physical impairment. Firstly, a dynamic impairment (DI) evaluation model is established to assess the level of impairment influence in terms of dynamic spectrum occupancy. Then, a DI-aware routing, modulation, core, and spectrum allocation (RMCSA) algorithm is proposed, which considers the influence of new lightpath establishment on existing lightpaths. Here, the core and spectrum resources with the least impact of impairment on existing lightpaths are selected preferentially to improve the QoT of lightpaths. Simulation results indicate that the proposed DI-aware RMCSA algorithm reduces the accumulated physical impairment on existing lightpaths, providing a lower probability of blocking and greater spectrum utilization.

Lightpath QoT computation in optical networks assisted by transfer learning

Precise computation of the quality of transmission (QoT) of lightpaths (LPs) in transparent optical networks has techno-economic importance for any network operator. The QoT metric of LPs is defined by the generalized signal-to-noise ratio (GSNR), which includes the effects of both amplified spontaneous emission noise and nonlinear interference accumulation. Generally, the physical layer of a network is characterized by nominal values provided by vendors for the operational parameters of each network element (NE). Typically, NEs suffer a variation in the working point that implies an uncertainty from the nominal value, which creates uncertainty in the GSNR computation and requires the deployment of a system margin. We propose the use of a machine learning agent trained on a dataset from an in-service network to reduce the uncertainty in the GSNR computation on an unused sister network, based on the same optical transport equipment and thus following the transfer learning paradigm. We synthetically generate datasets for both networks using the open-source library GNPy and show how the proposed deep neural network based on TensorFlow may substantially reduce the GSNR uncertainty and, consequently, the needed margin. We also present a statistical analysis of the observed GSNR fluctuations, showing that the per-wavelength GSNR distribution is always well-approximated as Gaussian, enabling a statistical closed-form approach to the margin setting.

Assessment on the in-field lightpath QoT computation including connector loss uncertainties

Reliable and conservative computation of the quality of transmission (QoT) of transparent lightpaths (LPs) is a crucial need for software-defined control and management of the wavelength division multiplexing optical transport. The LP QoT is summarized by the generalized SNR (GSNR) that can be computed by a QoT estimator (QoT-E). Within the context of network automation, the QoT-E must rely only on data from the network controller or provided by network elements through common control protocols and data structures. Therefore, given the theoretical accuracy of the QoT-E, the in-field accuracy in the GSNR computation is also determined by the level of knowledge of input parameters. Among these, a fundamental value is the connector loss at the input of each fiber span, which defines the actual power levels triggering the nonlinear effects in the fiber, and so defining the amount of nonlinear interference and spectra tilt due to the stimulated Raman scattering introduced by the fiber span. This value cannot be easily measured and may vary in time because of equipment update or maintenance. In this paper, we consider a lab measurement campaign in which the GSNR has been computed by means of the open source project Gaussian noise model in Python (GNPy) and analyze the computation error distribution. We show how the assumption on the value for the connector loss modifies the GSNR computation error and how the GSNR computation is more conservative while accurate at the lower values for the connector loss. Using the outcome of the measurement campaign carried out in the laboratory, we present results on the error of GSNR computation in a production network, specifically, over two paths of the Microsoft core network. Using GNPy with the assumption of a connector loss of 0.25 dB as derived from the measurement campaign carried out in the laboratory, and using the physical layer description from the network controller, we show that GNPy is not conservative by overestimating the GSNR in only 5% of cases, while in conservative predictions, the underestimation error exceeds 1 dB only for a few outliers.

Machine-Learning Method for Quality of Transmission Prediction of Unestablished Lightpaths

Predicting the quality of transmission (QoT) of a lightpath prior to its deployment is a step of capital importance for an optimized design of optical networks. Due to the continuous advances in optical transmission, the number of design parameters available to system engineers (e.g., modulation formats, baud rate, code rate, etc.) is growing dramatically, thus significantly increasing the alternative scenarios for lightpath deployment. As of today, existing (pre-deployment) estimation techniques for light-path QoT belong to two categories: “exact” analytical models estimating physical-layer impairments, which provide accurate results but incur heavy computational requirements, and margined formulas, which are computationally faster but typically introduce high link margins that lead to underutilization of network resources. In this paper, we explore a third option, i.e., machine learning (ML), as ML techniques have already been successfully applied for optimization and performance prediction of complex systems where analytical models are hard to derive and/ or numerical procedures impose high computational burden. We investigate a ML classifier that predicts whether the bit error rate of unestablished lightpaths meets the required system threshold based on traffic volume, desired route, and modulation format. The classifier is trained and tested on synthetic data and its performance is assessed over different network topologies and for various combinations of classification features. Results in terms of classifier accuracy are promising and motivate further investigation over real field data.

Hybrid Modulation Formats Enabling Elastic Fixed-Grid Optical Networks

In this paper, we analyze hybrid modulation formats as an effective technology for the implementation of flexible transponders that are capable of trading off the delivered data rate by the lightpath quality of transmission with fine granularity. Flexible transponders are an enabling technology that can introduce the elastic paradigm in state-of-the-art networks while maintaining compatibility with the installed equipment, including fibers, mux-demux, and reconfigurable optical add-drop multiplexers, as required by telecom operators willing to exploit fixed-grid wavelength-division multiplexed (WDM) transmission. We consider two solutions achieving different levels of flexibility and employing different hybridization approaches: time-division (TDHMF) and quadrature-division (Flex-PAM) hybrid modulation formats. We introduce a comprehensive theoretical assessment of back-to-back performances, analyzing different transmitter operating conditions, and we provide an extensive simulation analysis on the propagation of a Nyquist-WDM channel comb over an uncompensated and amplified fiber link. After assessing the impact of nonlinear propagation on the maximum signal reach, we present simple countermeasures for non-linear mitigation and discuss their effectiveness for both TDHMF and Flex-PAM.

Potentialities and Criticalities of Flexible-Rate Transponders in DWDM Networks: A Statistical Approach

We propose a novel method to assess physical layer potentialities of core optical networks aimed at finding solutions that better exploit the installed equipment. We focus on the use of flexible-rate transponders for the implementation of the elastic paradigm on the state-of-the-art dense wavelength-division-multiplexed fixed-grid network scenarios. We make use of the waveplane-based routing and wavelength assignment algorithm presented in Dai et al. [J. Lightwave Technol., vol. 33, p. 3815, 2015] to implement a progressive statistical loading of the analyzed network topology, and perform a Monte Carlo analysis delivering a statistical characterization of the average bit rate per lightpath together with the assessment of network blocking. The proposed method allows for the identification of criticalities in terms of link congestion and lightpath quality of transmission, addressing solutions by identifying network bottlenecks. We apply the proposed method to a large pan-European network topology comparing two different transmission techniques for the implementations of flexible-rate transponders: pure PM-m-QAM versus hybrid modulation formats. Over this realistic network example, besides displaying the overall statistics for the average bit rate per lightpath, we show statistics for critical lightpaths and congested fiber links.

Wavelength Preference in GMPLS-controlled Wavelength Switched Optical Networks

Given the potentialities in terms of high bandwidth, low costs, and low power consumption, Wavelength Switched Optical Networks (WSONs) are the most promising candidate for next generation backbone networks. In WSONs the optical signal is switched at the wavelength granularity, therefore the wavelength assignment process, selecting the carrier of each end-to-end optical connection (i.e., lightpath), plays a crucial role in dynamic network operation. Generalized Multi-Protocol Label Switching (GMPLS) is the standard control plane for WSONs, enabling dynamic lightpath provisioning and maintenance. However, current GMPLS protocol suite does not envision a general mechanism to rank the wavelengths candidate for carrying a lightpath. The lack of wavelength preference in WSONs may cause high blocking probability, wavelength converter waste, and detrimental effects due to physical impairments. This paper reviews several WSON scenarios where the wavelength preference concept is introduced to optimize the wavelength assignment: wavelength continuous, wavelength convertible, and quality of transmission aware WSONs. Automatic ranking of candidate wavelengths is effective to select the most suitable one in each considered scenario. To enforce wavelength preference, an extension to the GMPLS signaling protocol is utilized. Simulation results show that wavelength preference can effectively reduce blocking probability, save wavelength converters, and guarantee lightpath quality of transmission.