Off-line temporal analysis of hard real-time applications

Abstract

The most critical part in the development of hard real-time applications is the priority mapping carried out with nonformal methods. Our approach is based on the view that if the real-time application software is transformed into a set of tasks modeled by proper temporal characteristics, then the schedule analysis is easier to be determined. We propose simple environment models (language and real-time kernel) and a methodology of temporal modeling of the application starting from initial task configuration and going through to a set of periodic tasks sharing critical resources. This configuration allows us to make off-line timing proofs by using well-known scheduling algorithms based either on static priority like rate monotonic or on dynamic priority like earliest deadline. The resource requirements of tasks are performed with the priority ceiling scheduling protocol. The proposed temporal analysis permits us to evaluate different temporal properties of the real-time application.<>