Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Verifying network performance of cyber-physical systems with multiple runtime configurations

: Manderscheid, Martin; Weiß, Gereon; Knorr, Rudi

Postprint urn:nbn:de:0011-n-3642135 (287 KByte PDF)
MD5 Fingerprint: dcc0fa2e6ebbcaedfbca18f69ad60782
© IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Erstellt am: 12.11.2015

Institute of Electrical and Electronics Engineers -IEEE-:
EMSOFT 2015, 15th International Conference on Embedded Software. Proceedings : October 04-09, 2015, Amsterdam, Netherlands
Piscataway, NJ: IEEE, 2015
ISBN: 978-1-4673-8079-9
International Conference on Embedded Software (EMSOFT) <15, 2015, Amsterdam>
Konferenzbeitrag, Elektronische Publikation
Fraunhofer ESK ()
in-vehicle communication; feature model; mixed-integer linear programming; runtime variability; network performance; evaluation; cyber physical system; embedded system; automotive software; adaptive systems; reliable ethernet and IP communication

Modern Cyber-Physical Systems (CPS) must increasingly adapt to changing contexts, like smart cars to changing driving conditions. Thus, design approaches are facing a rapidly growing number of network runtime configurations. With recent approaches this problem can be solved for design space exploration (DSE) by analyzing the network performance of single configurations which are intended to represent the entire runtime variability space. This technique can be applied for DSE since the latter only intends to find an optimized system setup. Yet it does not meet the requirements of network verification, since it does not necessarily find the worst-case for all applications. To solve this, we developed an integrated model, which allows describing runtime variability in the network performance model with a0-1 linear-fractional program. Thus, we can cover entire runtime variability spaces without analyzing every single network runtime configuration. Although the approach utilizes heuristics, it still guarantees worst-case results. We can show that in comparison to state-of-the-art methods our approach scales for large automotive systems with multiple network configurations. Moreover, our evaluation results highlight the superior capabilities of our method with respect to accuracy and computation time.