Analyzing the Timing Characteristics of Task Activations

Abstract

We propose a model and method to determine the temporal distribution of task activations occurring within an event driven real time system. As real time systems even in safety critical applications are becoming increasingly complex, it becomes increasingly important to get a better understanding of the activity within these systems. The event stream model is a powerful yet efficient way to describe the temporal occurrence of events within a system. Real time analysis can be elegantly performed using given event streams, however it has so far not been investigated how the event streams themselves can be acquired. By calculating the temporal distribution of task activations within a system, a more accurate description of the timing behavior of a system is possible which in the end can lead to a save and cost efficient estimate on the hardware required for the system.