White Rabbit ns Time Transfer
Summary
2. Safe and Reliable Systems
Armasuisse
Armasuisse; METAS
Jacques Robadey
Skills directory
June 2021 - February 2022
Development of a demonstrator for time distribution from a reference time with high accuracy (sub ns) as a service for users of operational highspeed IP networks (Switch/FUB).
This project aims to develop a demonstrator for time distribution from a reference time with high accuracy (sub ns) as a service for users of operational highspeed IP networks (Switch/FUB). The White Rabbit technology developed at CERN and marketed by different vendors should allow for achieving this performance. The chosen loop topology will have to provide the reference time to two users from two independent time masters localised at the same place and provide redundancy in case of network component failure.
The project was financed by Armasuisse and performed in close collaboration with METAS. A national network demonstrator using the White Rabbit’s (WR) technology has been built at the HEIA-FR over 200km of fibres to synchronize distant clocks with sub ns accuracy. After calibration of the WR devices, we performed a detailed analysis of the system including daisy-chain ring topology to achieve reliable time distribution and high synchronization accuracy.
The tests of multiple WR devices and network loops gave those results: the media converters used as optical/electrical/optical signal repeaters were assessed to fit the project’s needs. Sub ns accuracy specification has been achieved with a global standard deviation of 50 ps. We observed high system’s stability and reliability over time with continuous testing for 2-3 days showing no abnormal or inconsistent behaviour. The network is stable when the temperature of Small Form-Factor Pluggable (SFP) devices is monitored and compensated. Otherwise, tests show a time variation of 2.6 ps/km/°C when a slave SFP is heating up. Overall results comply with the project’s expectations.
After the demonstration at HEIA-FR, the system has been installed in a real network covering four times the distance “Thun-Bern” distance and showing a synchronization accuracy of 5ps.
Project's team
Jacques Robadey - project manager
Pamela Torche - scientific collaborator
Florian Helfer - scientific collaborator