a predictable execution strongmodelstrong for cots-based embedded strongsystemsstrong.pdfVIP

A Predictable Execution Model for COTS-based Embedded Systems Rodolfo Pellizzoni , Emiliano Betti , Stanley Bak , Gang Yao , John Criswell , Marco Caccamo and Russell Kegley University of Waterloo, Canada, rpellizz@uwaterloo.ca University of Illinois at Urbana-Champaign, USA, ebetti, sbak2, criswell, mcaccamo@ Scuola Superiore Sant’Anna, Italy, g.yao@sssup.it Lockheed Martin Corp., USA, russell.b.kegley@ Abstract—Building safety-critical real-time systems out of triple. To exploit the high average performance of COTS inexpensive, non-real-time, COTS components is challenging. components without occasionally experiencing long delays Although COTS components generally offer high performance, suffered by real-time tasks, we need to control the operating they can occasionally incur significant timing delays. To prevent this, we propose controlling the operating point of each shared point of each shared resource and maintain it below saturation resource (like the cache, memory, and interconnection buses) to limits. This work aims at showing that this is indeed possible maintain it below its saturation limit. This is necessary because by carefully rethinking the execution model of real-time tasks the low-level arbiters of these shared resources are not typically and by enforcing a high-level coscheduling mechanism among designed to provide real-time guarantees. In this work, we all active components in the system. Briefly, the key idea introduce a novel system execution model, the PRedic