https://www.mdu.se/

The new generation networks and systems are increasingly demanded to present higher capacities such as flexibility, adaptability, or real-time capabilities. In this context, Time-Sensitive Networking (TSN) standards provide Ethernet with functionalities capable of meeting these new demands. However, the adoption of TSN by the industry has been limited due to the high complexity of configuring and analyzing its various mechanisms. This thesis aims to address this issue by developing tools to facilitate the adoption of TSN in the industry. To this end, we have created tools to enable efficient and semi-automatic TSN configuration and analysis. These tools can efficiently configure TSN networks based on traffic characteristics while supporting legacy devices in case of migration or integration of legacy systems to TSN. Additionally, the analysis tools enable guaranteeing the compliance of the TSN configuration with the network requirements with reduced waste of resources.

Specifically, we have devised a mapping tool called LETRA to distribute the Ethernet-based traffic among the three basic types of TSN traffic, i.e. Scheduled Traffic (ST), Audio-Video Bridging Traffic (AVB), and Best-Effort Traffic (BE). Since ST requires traffic scheduling, we developed HERMES, a heuristic scheduler that generates efficient schedules in a fast manner. For AVB traffic, which requires some previous schedulability analysis to guarantee it meets its time requirements, we propose a new Worst-Case Response Time Analysis (WCRTA) which reduces the pessimism of previous approaches. Finally, regarding the support for the integration of legacy devices in TSN systems, we have developed mechanisms to avoid adverse consequences resulting from the lack of synchronization between legacy devices and TSN.