Modeling and controlling optical transponder white-boxes based on the physical layer digital twin

The growing demand for high data rates and the increasing popularity and capillary of high-speed access, both wired and for mobile use, have made it clear that a robust and reliable core data network is essential. The improvement of the network performances can be obtained in several ways; however, a crucial aspect is defining an open way of controlling all network devices, exploiting multi-vendor setups in an innovative and efficient way. This is a first step towards the realization of a digital twin, or a software representation of a real network, which can be used to make simulations and to plan improvements. This is the foundation of the initial section of this work, which examines the potential of an open approach to control a popular whitebox transponder that is compatible with commercial transceivers. These pluggable devices are at the basis of the optical networks, since they are the elements in charge of converting the signals between the electrical and the optical domain. In recent years, their internal implementation has improved significantly, enabling their use in a vast set of operating conditions while also improving their flexibility at operating at different frequencies, rates, and modulation formats. This has led to the development of extremely high-speed products, reaching data rates of up to 400 Gbps or beyond. This lead to the need of defining and implementing a new open driver able to exchange configuration parameters and statistical data between proprietary interfaces typical of the used transponder and an open standard that can be easily integrated into a more complex network topology control software. In this way, it has been shown how the main characteristics of the optical transceivers are made available to be controlled in an open and flexible way. In the second part of this work, the newly developed open control is exploited in conjunction with the use of other devices in a laboratory environment, including optical amplifiers, attenuators, noise generators, and optical spectrum analyzers, to perform a comprehensive characterization of the said transponder. In practice, the use of these devices is of great importance in order to perform a large data collection of several important parameters of the device's behavior in different operating conditions. Initially, the setup has been used in order to measure the limits of the device, and to understand which are its optimal working conditions, and which are the penalties when the devices are used outside that region. Furthermore, a detailed analytical study of the internal device components has been conducted to mathematically define the source of internal noise at the transmitter and receiver, which, along with the amplifiers, represent the primary source of noise in a lightpath. Finally, the joint use of the mathematical model and the experimental results enabled the possibility of defining a complete model able to accurately predict the performances of the device with the help of some key parameters, which can be directly interpolated from the data collection itself. In conclusion, the work establishes a solid foundation for additional analysis that can be used to further improve the model or tailor it to different devices and user goals. This makes it an open solution and starting point for the definition of a complete open network design, which is at the basis of the realization of a full network digital twin.