The Physical Simulation Environment (PSE) is an initiative within the Telecom Infra Project (TIP), focusing on the development and deployment of innovative solutions for disaggregated optical networks. Its major contribution is the development of GNPy, an open source Optical Route Planning and DWDM Network Optimization based on Gaussian Noise modelling in Python. PSE brings together experts from academia, industry, and telecommunications to explore advanced methods for optimizing network performance, improving control mechanisms, and leveraging digital twins in real-world network management.
A key contribution by the PSE subgroup this year is a new scientific article titled “QoT Digital Twin for Bridging Physical Layer Knowledge Gaps in Multi-Domain Networks,” which has been accepted for oral presentation at Optical Fiber Conference (OFC) 2025. This work, produced by Politecnico di Torino in collaboration with NTT Network Innovation Labs, Orange, Telia Company, NEC Laboratories America Infinera and Juniper among others, explores the use of digital twin models and transponder telemetry to bridge the knowledge gaps in the physical layer of multi-domain optical networks. Specifically, it proposes a methodology to infer the optical performance of a network domain with limited available information, enabling more efficient management and operation of disaggregated networks.
Why is this important?
This paper presents a solution to optimize multi-domain optical networks with incomplete information about optical network characteristics by leveraging digital twin models. As networks become increasingly complex, managing different segments with incomplete information is a significant challenge. This work contributes to TIP OOPT’s efforts, particularly within the MUST use cases, by enhancing the planning and management of multi-domain networks. By integrating real-time telemetry data and performance models, the paper offers insights for improving GNPy simulations, enabling efficient network planning, improved Quality of Transmission (QoT) assessments, and optimized resource management across domains. This research supports TIP OOPT’s vision for an open, flexible, and software-controlled optical transport ecosystem.
The paper introduces a digital twin (DT) model that combines transponder telemetry data with a digital representation of the optical network. This model helps to fill-in information in domains with limited visibility i.e. information that cannot be retrieved from the network itself, such as amplifier properties and fiber characteristics. The target is to improve QoT prediction and facilitate the deployment of higher bit-rate services, in network areas with incomplete visibility of optical characteristics.
OFC Presentation Details
- Presenting Author: Renato Ambrosone
- Session Abbreviation: W4G
- Session Title: Digital Twins in Network Control and Management
- Date and Time: April 2, 2025, from 4:30 PM to 6:30 PM PDT
- Location: Rooms 213-214 (Level 2)
This research is a collaboration among multiple institutions and industry partners:
Authors: Renato Ambrosone¹, Andrea D’Amico², Riccardo Schips¹, Andrea Rosso¹, Esther Le Rouzic³, Stefan Melin⁴, Stefano Straullu⁵, Giacomo Borraccini², Francesco Aquilino⁵, Hideki Nishizawa⁶, Sai Kishore Bhyri⁷, Gert Grammel⁸, and Vittorio Curri¹.
¹Politecnico di Torino, Turin, Italy, ²NEC Laboratories America Inc., Princeton, USA., ³Orange, Innovation Networks, Lannion, France. ⁴Telia Company AB, Solna, Sweden., ⁵LINKS foundation, Turin, Italy., ⁶NTT Network Innovation Labs, Kanagawa, Japan., ⁷Infinera India Private Limited, India., ⁸Juniper Networks, Munchen, Germany.
We look forward to discussing our findings and engaging with the community at OFC 2025. If you are attending, we welcome you to join our session and exchange insights on the future of digital twins in optical network management.