Holistic scheduling and analysis of mixed time/event-triggered distributed embedded systems

Abstract

This paper deals with specific issues related to the design of distributed embedded systems implemented with mixed, eventtriggered and time-triggered tusk sets, which communicate over bus protocols consisting of both static and dynamic phases. Such systems are emerging as the new standard for automotive applications. We have developed a holistic timing analysis and scheduling approach for this category of systems. We have also identified several new design problems characteristic to such hybrid systems. An example related to bus access optimization in the context of a mixed static~dynamic bus protocol is presented. Experimental results prove the efficiency of such an optimization approach.

Composing Functional and State-Based Performance Models for Analyzing Heterogeneous Real-Time Systems

“…Before concluding this section, we would like to point out that our approach is orthogonal to the previous efforts on full-system performance analysis, for example, using holistic schedulability analysis [2], [12], [13], [14], [17]. Although these efforts were also directed towards analyzing systems with multiple resources and scheduling policies, the underlying analysis techniques used for the different components were similar and are mostly based on classical event models.…”

Analysis and Optimisation of Hierarchically Scheduled Multiprocessor Embedded Systems

“…The first part consists in a holistic schedulability analysis for heterogeneous TT/ET task sets which communicate through mixed ST/DYN communication protocols [PopT02], [PopT03a].…”

Dataflow Analysis for Real-Time Embedded Multiprocessor System Design

“…The typical temporal behavior of jobs depends on the values of the input data which are unknown at compile time. Therefore, purely model based analysis techniques for hard real-time jobs, such as the techniques in (Kopetz, 1997;Pop, Eles and Peng, 2002;Richter, Jersak and Ernst, 2003), are not directly applicable for the analysis of soft real-time jobs. The reason is that the actual values of the input data can be ignored during analysis of hard real-time jobs because the obj b ective is to derive the worst-case temporal behavior for any possible input data stream.…”