Dependencies aware event-driven real-time analysis for distributed fixed-priority systems

In this paper we present an approach to calculate the maximum density of events in a distributed hard real-time system having tree-shaped dependencies. Thereby we will present how it is possible to relax the density of events in such a system by including scheduling dependencies. This relaxation has a direct impact of successive tasks and leads to more realistic real-time analysis. In this paper we distinguish between the task model and the model for the stimulation with the result that we can describe a major range of stimulation. In the end we will show how it is possible to make a real-time analysis with the presented approach.