Optimization on Time-Sensitive Networking in Avionic Systems Industries, especially the aircraft industry heavily rely on real-time communication and for that reason highly specific network technologies were developed to fulfil those requirements. This specialization reduces the capability to use commercial off-the-shelf (COTS) component, that leads to higher costs. Thus Time-Sensitive Networking (TSN) can provide deterministic, low-latency, and high reliability transmission. With these properties it can act as the central network in real-time systems - like in Industrial Automation, Automotive, and Avionics Systems - creating a backbone to connect different communication protocols and allow a high control over data traffic with the usage of Traffic Shapers. TSN is still a growing standard that aim to utilize bandwidth while still providing reliable communication with redundancy, precise time synchronization and other strategies to have accurate and timely data transmission. Utilizing TSN increases the complexity of traffic management and the aim for automatized configuration creates new challenges such as the defining Traffic Shaper parameters and creating Multi-Shaper configurations.

Traffic Shapers are a key feature of TSN as the utilization provides better performance and higher reliably. However, the complexity of the configuration and parameter settings of the Traffic Shaper depends on multiple factors and varies accordingly from straightforward to highly complex. This also means that it depends highly on the network definition, requirements, the amount of nodes, amongst others and the used Traffic Shaper(s) themselves. Manual configuration is time consuming and does not support flexible systems, changing network configurations, and high level automatization. The configuration parameters being highly dependent on the network create major automatization and optimization opportunity.

All of the above raises the question on how to select and combine the different Traffic Shapers for the best result. And how to automate generation of Traffic Shaper parameters for automatic (re)configuration on self adaptive aviation systems.

The general approach for Traffic Shaper implementation in low complexity systems can be solved with using pre-configured TSN switches and manual configuration. However, with more complex requirements configuring TSN is done simulating the network behaviour with accurate traffic volume monitoring with management tools. These management tools allow the parameter configuration and manual optimization based on the monitored performance of the simulated network.

The aim of this talk is to explore methods for automatic optimization of traffic shaping parameters based on Integer Linear Programming. This shall provide comparable results of Traffic Shaper parameter configurations and also for the performance differences of Multi-Shaper combinations with what the network configuration fulfils the requirements.