2020/04/20 by Meysam Goodarzi, Goodarzi, Meysam, Darko Cvetkovski +7
Computer Science · Engineering · #Advanced Optical Network Technologies #FOS: Electrical engineering #Network Time Synchronization Technologies #Signal Processing (eess.SP) #Software-Defined Networks and 5G #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2004.09469
openalex publication_date 2020/04/20 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
In this work, we propose a hybrid Bayesian approach towards clock offset and\nskew estimation, thereby synchronizing large scale networks. In particular, we\ndemonstrate the advantage of Bayesian Recursive Filtering (BRF) in alleviating\ntime-stamping errors for pairwise synchronization. Moreover, we indicate the\nbenefit of Factor Graph (FG), along with Belief Propagation (BP) algorithm in\nachieving high precision end-to-end network synchronization. Finally, we reveal\nthe merit of hybrid synchronization, where a large-scale network is divided\ninto local synchronization domains, for each of which a suitable\nsynchronization algorithm (BP- or BRF-based) is utilized. The simulation\nresults show that, despite the simplifications in the hybrid approach, the Root\nMean Square Errors (RMSEs) of clock offset and skew estimation remain below 5\nns and 0.3 ppm, respectively.\n