Accelerating Nokia Optical innovations

Nokia’s (ex-Infinera/Coriant) represents the commercial evolution of a long history of innovation. The optical networking team specializes in transforming fundamental research into deployable solutions: scalable optical network platforms. Their commercial line of products is the backbone of today’s high-speed internet, pioneering foundational capabilities in dense wavelength division multiplexing (DWDM), submarine cable systems, high-capacity metro regional transport, and now, state of polarization tracking for sensing. Nokia’s team are committed to supporting global service providers to meet the relentless demand for bandwidth and network awareness by conducting research that creates sustainable and future-proof optical solutions.

Enabling Mass-Scale Optical Sensing through Telecom Innovation

The drive toward mass-scale sensing and Earth observation over optical networks will benefit from both application-specific systems – which deliver optimal accuracy and measurand fidelity – and enhanced telecom network equipment, enabling large-scale deployment, seamless coexistence and cost-efficiency.

Nokia expertise contribution for the SENSEI project is twofold. First, building on its long-standing expertise in compensating environmental effects through advanced signal processing for high-quality data transmission, Nokia enables network impairments traditionally treated as noise to be transformed into valuable network information. In particular, Nokia’s recent generations of transponders can stream State of Polarization performed at the receiver, allowing the detection and analysis of sudden disturbances along the optical link.

Second, Nokia contributes its expertise in data infrastructure and management, by implementing novel Machine Learning (ML) compression and synthetic data generation schemes. These approaches enable a practical and manageable system output, while also providing high-quality data training data for sparsely occurring phenomena.

Advanced Machine Learning for Telemetry Compression and Data Generation

Recent results obtained by Nokia’s research team comparing the ML-based data compression of telemetry using Autoencoder (AE) to the traditional approach on Principal Component Analysis (PCA) is highly relevant for SENSEI. Autoencoder shows more accurate reconstructed data with the same compression ratio and memory saved. This improved performance comes at the cost of increased model training complexity.

Building on these results, the project will move forward to using this machine learning approach based on variational AE (VAE) to generate synthetic data to address the imbalance dataset of the rare event. Nokia will also deploy dedicated hardware (Nokia 800G high-end optical system) to GARR network (Politecnico di Torina / GARR node) to support experimental validation.

Benchmarking Optical Network Awareness in Real-World Conditions

In SENSEI, Nokia’s team has the chance to benchmark its system’s current capabilities with other state-of-the-art approaches, and study them jointly towards a fully aware optical network that can serve the public. Nokia’s target is to execute at least two measurement campaigns in real optical networks with the same equipment and in places with very different seismicity: The Torino metropolitan area, with a low seismicity, hosted by Links and Politecnico di Torino, and in GARR’s network starting from Rome and crossing the Appenines, with a high seismicity. The outputs will provide an understanding of the current system’s capabilities and how to drive it forward for future enhanced performance.

…more to come! Stay tuned…