Abstract
To safely establish the correct timing of a real-time processing node, adequate architectural structures have to be used. This refers to the hardware architecture of the processing node as well as the software architecture of its operating system and application software. This paper presents architectures that allow for a well structured and simple timing analysis. First, it presents solutions for cleanly splitting the overall timing analysis into schedulability analysis and task worst-case execution time analysis. Second, it presents a programming strategy that yields software that is highly temporally predictable and easy to analyze for its worst-case execution time.