In the context of Industry 4.0, achieving flexible and deterministic data delivery is crucial, particularly within networks that leverage Time-Sensitive Networking (TSN) -- an extension of traditional Ethernet -- for real-time service guarantees. Current TSN configurations, however, are predominantly static and do not offer the adaptability required to cater to the continuously evolving demands of applications. Adaptations become increasingly complex when changes need to be incorporated during operation, and applications demand more stringent service guarantees. This complexity stems from elaborate TSN planning requirements, resulting in static switch configurations. Furthermore, these configurations lack indicators for the reconfiguration potential, thus making and predicting adaptations difficult. The static nature of TSN schedule configurations typically necessitates complete reconfigurations, resulting in service downtimes for applications in the network.

To address these challenges, this thesis introduces a novel flexibility metric called "flexcurve" designed to enhance the management of dynamic TSN networks. The introduction of the flexcurve metric aims to quantify the flexibility at network bottlenecks by considering frame size and other traffic requirements, thereby measuring possibilities for the integration of new traffic without extensive rescheduling. Consequently, it enables a central TSN controller to perform more informed traffic admissibility checks, path selection, and scheduling adjustments during the configuration planning phase. Utilizing a path-based approach, the flexcurve metric captures the essence of end-to-end scheduled traffic. Furthermore, the formulation allows for both disaggregation and aggregation, enhancing efficiency. Disaggregations enable traffic admissibility checks of different sources simultaneously, while aggregations facilitate the rapid construction of the metric from existing configurations.

The thesis builds on the flexcurve metric and introduces a search heuristic algorithm that employs the flexcurve for dynamic scheduling. This algorithm is designed to generate eligible candidates and incorporates a secondary heuristic for pruning to select the most promising candidates. Additionally, methods are included for choosing paths that optimize the value of the flexcurve.

As final contributions, the deployment capabilities of TSN to non-TSN hardware have been augmented by leveraging programmable Push-In-First-Out (PIFO) queues and by developing the Residence Delay Aggregation (RDA) method, which ensures per-flow delay guarantees on programmable switches. Additionally, the flexcurve has been further extended by integrating multi-mechanism support, significantly improving network flexibility and broadening TSN deployment opportunities.