Browsing by Author "Philipp, Lars"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Investigation and verification of real-time schedules via simulation
    (2024) Philipp, Lars
    With the rise in popularity of Cyber-Physical Systems (CPSs) in various domains such as the Internet of Things (IoT) and autonomous vehicles, the demand for real-time systems especially with deterministic real-time communication, has significantly grown. Bounded latencies are a crucial part of the safety of CPSs, maintained through deterministic real-time communications. The Institute of Electrical and Electronics Engineers (IEEE) recognized the rise in popularity of such real-time systems with deterministic communication, resulting in a set of standards known as Time- Sensitive Networking (TSN). Although this standard establishes the technological infrastructure for deterministic communication on Ethernet networks, it does not specify the computation of schedules for TSN networks. This led to the development of scheduling algorithms for TSN networks becoming a major field of research. A critical step prior to the deployment of these algorithms is to investigate and verify them. In this study, we investigate and verify the scheduling algorithms through a simulation of the computed schedule on TSN networks. A simulation offers an easy and cost-effective way to verify the scheduling algorithms. We utilize the discrete event simulator OMNeT++ (Objective Modular Network Test-bed in C++) with the INET framework for the simulation of the scheduling algorithms. In particular, we simulate schedules generated with the Greedy Flow Heap (GFH) algorithm. This algorithm uses a conflict-graph-approach to compute the schedules. In order to facilitate the simulation process, we have developed a pipeline that effectively transforms the computed schedules of the algorithm into a fully operational simulation. In our evaluation, we successfully compared the computed schedules of the scheduling algorithm with the simulation results. Furthermore, we show the limitations of the scheduling algorithm, which are primarily attributable to the existing implementation of the algorithm.