Time-Predictable Software-Based TSN-Enabled Network Stack for Mixed Criticality Traffic

Abstract

The ongoing transition in industrial automation towards a highly interconnected architecture underscores the importance of time-sensitive networking (TSN). This paper addresses the implementation of a time-predictable software-based TSN-enabled network stack for mixed criticality traffic in TSN end stations. The findings demonstrate the network stack's worst-case execution time (WCET) analysability, effective management of mixed critical traffic, and real-time transmission of UDP frames. Additionally, the stack offers modularity, configurability, and TSN support for transferring messages across different priorities using either UDP or TCP. Evaluation on the time-predictable open-source T-CREST platform yields representative WCET estimates comparable to related research. The paper concludes by outlining further research.