A Model-Based Approach to Automatic Generation of TSN Network Simulations

• Published in: 2022 IEEE 18th International Conference on Factory Communication Systems (WFCS) pp. 1 - 8
• Main Authors: Samson, Maxime, Vergnaud, Thomas, Dujardin, Eric, Ciarletta, Laurent, Song, Ye-Qiong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 27.04.2022
• Subjects: Complexity theory; Configuration generation; Data models; Ethernet; MARTE specification; Model-Based software; Real-time systems; Simulation; Software; Switches; TSN; Writing
• DOI: 10.1109/WFCS53837.2022.9779173

The IEEE 802.1 TSN working group published a set of standards which adds new functionalities to switched Ethernet networks. These new functionalities notably aim at making the design of deterministic Ethernet networks possible, enabling their use for real-time applications. This determinism comes however at the cost of a greatly increased complexity in configuration effort, especially for largescale networks. In addition, this new complexity also affects the network simulation tools that are commonly used when designing such networks. As most of the existing ones only partially support new TSN functionalities, one has often to combine several simulators for achieving a reliable TSN network design. In this paper, we propose a model-based approach which aims at assisting the design of TSN networks. Modeling allows the creation of a formal representation of the network which can then be used to automatically generate configurations. This approach has been successfully implemented as a TSN configuration software called MoBACT, that enables the use of multiple simulation/emulation tools during the design phase by generating configurations for different targets for easing crosscheck of simulation results. Since each generated configuration is derived from the same representation of the network, our approach guarantees the consistency of the different configurations generated for each tool.