A survey on QoT prediction using machine learning in optical 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.

Recently, machine learning (ML) has attracted the attention of both researchers and practitioners to address several issues in the optical networking field. This trend has been mainly driven by the huge amount of available data (i.e., signal quality indicators, network alarms, etc.) and to the large number of optimization parameters which feature current optical networks (such as, modulation format, lightpath routes, transport wavelength, etc.). In this paper, we leverage the techniques from the ML discipline to efficiently accomplish the routing and wavelength assignment (RWA) for an input traffic matrix in an optical WDM network. Numerical results show that near-optimal RWA can be obtained with our approach, while reducing computational time up to 93% in comparison to a traditional optimization approach based on integer linear programming. Moreover, to further demonstrate the effectiveness of our approach, we deployed the ML classifier into an ONOS-based software defined optical network laboratory testbed, where we evaluate the performance of the overall RWA process in terms of computational time.

Impairment aware routing and wavelength assignment algorithms for optical WDM networks and experimental validation of impairment aware automatic light-path switching

Transparent All Optical Networks carry huge traffic and any link failure can cause the loss of gigabits of data; hence protection and its guarantee becomes necessary at the time of failure. Many protection schemes were presented in the literature, but none of them speaks about protection guarantee. Also, in all optical networks, due to absence of regeneration capabilities, the physical layer impairments(PLI) accumulates along the lightpaths (LP) which causes sharp degradation of the Quality of Transmission (QoT), as measured by signal bit error rates (BER). The problem of protection with QoT issues was rarely studied in the literature. In this work, a novel protection backup path ensured, QoT aware Routing and Wavelength Assignment (RWA) algorithm called “Virtual Lit –Exhaustive Highest Q factor” (V-Lit EHQ) is presented which exhibits desirable qualities for reliable network operation. The results of the proposed work are compared with the standard QoT aware versions of the Shortest Path-First Fit schemes for both lit and dark protection. The blocking probability (BP) and BER are taken as the performance matrices and the proposed algorithm found to be outperforming as shown through simulations.

In translucent optical networks, the physical layer impairments degrading the optical signal are considered in the network planning. In this paper, we investigate the offline problem of routing and wavelength assignment (RWA) and regenerator placement (RP) in translucent networks, minimizing the lightpath blocking and regenerator equipment cost. We address two variants of the problem, which correspond to two different types of quality of transmission (QoT) estimators, called linear and nonlinear. In a nonlinear QoT, nonlinear impairments like crosstalk or cross-phase modulation, which account for the interferences from neighboring lightpaths in the network are explicitly computed. Then, the QoT estimated for a lightpath depends on the routes of other lightpaths in the network. In the linear QoT, the effects of the nonlinear impairments are overestimated and accumulated to the rest of the impairments in the QoT calculation. As a result, the QoT estimation of a lightpath solely depends on its route. For the linear case, we formulate an optimal integer linear programming model of the problem, to the best of the authors' knowledge, for the first time in the literature. Its simplicity allows us to test it for small- and medium-size networks. Also, we propose two heuristic methods, namely, lightpath segmentation and three-step, and a tight lower bound for the regenerator equipment cost. For the nonlinear QoT case, we propose a new heuristic called iterative RP (IRP). Both the IRP and three-step algorithms are designed to guarantee that no lightpath blocking is produced by signal degradation. This is a relevant difference with respect to earlier proposals. The performance and the scalability of our proposals are then investigated by carrying out extensive tests. Results reveal that the solutions obtained by the heuristic algorithms are optimal or close to optimal, and outperform the earlier proposals in the literature.

WDM Network Issues: QOT Impairments, Dynamic Bandwidth Allocation, Modeling and Monitoring Schemes based on ML

Presently the demand from users for communication networks with high bandwidth, high data rate, high security, and reliable connections has suddenly surged due to increased internet applications over the last five years. This is making communication networks complex. Further use of optical fiber cables as the backbone because of their high bandwidth and capacity, which can carry Tera bits of information, has further increased the complexity of accessing and analyzing the optical networks due to their non-linear behavior. In this regard, the availability of advanced machine learning algorithms based on mathematical tools, which are in a mature stage, has been looked upon. Machine learning is becoming an important tool because of the unexpected growth of network complexity. In the optical network, resource efficiency and capacity of the network is determined by routing and resource allocation. Since optical networks are the backbone, making them energy efficient will reduce carbon footprints and hence can reduce global warming. Hence, mathematical tools of machine learning algorithms are explored for optical communication systems and networks. Applications such as nonlinear equalization, optical performance monitoring, signal detection, mitigating various noise effects etc., and corresponding to their machine learning tools are discussed in detail. Researchers can explore further using machine learning tools in an optical domain and hence can propose and enhance flexibility and reliability.

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.

Machine learning (ML) is increasingly applied in optical network management, especially in cross-layer frameworks where physical layer characteristics may trigger changes at the network layer due to transmission performance measurements (quality of transmission, QoT) monitored by optical equipment. Leveraging ML-based QoT estimation approaches has proven to be a promising alternative to exploiting classical mathematical methods or transmission simulation tools. However, supervised ML models rely on large representative training sets, which are often unavailable, due to the lack of the necessary telemetry equipment or of historical data. In such cases, it can be useful to use training data collected from a different network. Unfortunately, the resulting models may be uneffective when applied to the current network, if the training data (the source domain) is not well representative of the network under study (the target domain). Domain adaptation (DA) techniques aim at tackling this issue, to make possible the transfer of knowledge among different networks. This paper compares several DA approaches applied to the problem of estimating the QoT of an optical lightpath using a supervised ML approach. Results show that, when the number of samples from the target domain is limited to a few dozen, DA approaches consistently outperform standard supervised ML techniques.

Machine learning framework for predicting failure mode and shear capacity of ultra high performance concrete beams

With the rapid advancement of information technology, business intelligence and data analysis have become integral to the success of modern enterprises. Through the utilization of various computing technologies, organizations can gain a better understanding of their vast amounts of data and optimize their business decisions. Machine learning technology, in particular, has presented significant potential for enterprises. By analyzing historical sales data and relevant factors, machine learning can predict the sales volume of goods. Commonly used algorithms in machine learning for this purpose include regression algorithms and neural network algorithms. By establishing mathematical models, future sales trends can be accurately forecasted. These algorithms have the capability to automatically identify hidden patterns and trends within the data, enabling enterprises to predict demand and adjust their production and supply chain strategies accordingly, thus better meeting market demand. By integrating regression algorithms and neural network algorithms with enterprise data, managers can obtain real-time and precise sales forecasts. These forecast results provide robust support for enterprise decision-making, enabling operators to formulate more informed marketing strategies, optimize inventory management, and make accurate production plans during peak demand periods. Additionally, by analyzing sales data and other relevant information, enterprises can identify key factors that influence sales, such as product characteristics, market trends, and consumer behavior. This deeper understanding of the market and customer needs allows organizations to enhance their competitiveness further. The application of machine learning technology, specifically regression algorithms and neural network algorithms, coupled with comprehensive data analysis, empowers enterprises to make more accurate sales forecasts. These forecasts aid in strategic decision-making, including marketing strategies, inventory management, and production planning, while also enabling businesses to gain valuable insights into market dynamics and customer preferences, ultimately strengthening their competitive advantage.

For reliable and efficient network planning and operation, accurate estimation of Quality of Transmission (QoT) before establishing or reconfiguring the connection is necessary. In optical networks, a design margin is generally included in a QoT estimation tool (Qtool) to account for modeling and parameter inaccuracies, ensuring the acceptable performance. In this article, we use monitoring information from an operating network combined with supervised machine learning (ML) techniques to understand the network conditions. In particular, we model the penalties generated due to i) Erbium Doped Fiber Amplifier (EDFA) gain ripple effect, and ii) filter spectral shape uncertainties at Reconfigurable Optical Add and Drop Multiplexer (ROADM) nodes. Enhancing the Qtool with the proposed ML regression models yields estimates for new or reconfigured connections that account for these two effects, resulting in more accurate QoT estimation and a reduced design margin. We initially propose two supervised ML regression models, implemented with Support Vector Machine Regression (SVMR), to estimate the individual penalties of the two effects and then a combined model. On Deutsche Telekom (DT) network topology with 12 nodes and 40 bidirectional links, we achieve a design margin reduction of ~1 dB for new connection requests.

P125. Development of a novel ensemble machine learning algorithm for prediction of complications and readmission after anterior cervical spinal fusion

Physical layer impairments, in Wavelength-Division Multiplexing (WDM) transparent networks, deteriorate the analog signal quality when it progresses from its source node to its destination node. Thus, the lightpaths' Quality of Transmission (QoT) might become potentially unacceptable. Such degradation has a severe impact on the performance of long-haul transmission transparent optical networks in terms of rejection ratio. This is especially severe when considering Random Lightpath Demands (RLDs). Traffic rerouting is a good solution that can improve transparent optical networks performances. This paper investigates rerouting and minimization of incurred service disruption period. Numerical experiments show that the proposed impairment-aware wavelength rerouting scheme provides significant improvement either in terms of rejection ratio or in terms of service disruption period.

Physical layer impairments, in Wavelength-Division Multiplexing (WDM) transparent networks, deteriorate the analog signal quality when it progresses from its source node to its destination node. Thus, the lightpaths' Quality of Transmission (QoT) might become potentially unacceptable. Such degradation has a severe impact on the performance of long-haul transmission transparent optical networks in terms of rejection ratio. This is especially severe when considering Random Lightpath Demands (RLDs). Traffic rerouting is a good solution that can improve transparent optical networks performances. This paper investigates rerouting and minimization of incurred service disruption period. Numerical experiments show that the proposed impairment-aware wavelength rerouting scheme provides significant improvement either in terms of rejection ratio or in terms of service disruption period.