Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Security overhead and its impact in VANETs

: Bittl, Sebastian; Roscher, Karsten; Gonzalez, Arturo A.

Postprint urn:nbn:de:0011-n-3642124 (177 KByte PDF)
MD5 Fingerprint: 0dd775ff1c03ed978780eddd5a32d7b6
Created on: 18.2.2016

Institute of Electrical and Electronics Engineers -IEEE-; International Federation for Information Processing -IFIP-:
8th IFIP Wireless and Mobile Networking Conference, WMNC 2015. Proceedings : 5-7 October 2015, Munich, Germany
Piscataway, NJ: IEEE, 2015
ISBN: 978-0-7695-5662-8
Wireless and Mobile Networking Conference (WMNC) <8, 2015, Munich>
Conference Paper, Electronic Publication
Fraunhofer ESK ()
cross layer optimization; security; VANET; vehicular ad hoc network; Car-to-x; Car2X; driver assistance; automotive connectivity

Vehicular ad hoc networks (VANETs), often called Car2X communication systems, are about to enter the mass market in upcoming years. They are intended to increase traffic safety by enabling new safety critical driver assistance systems. This also means that strong security mechanisms are required to safeguard communication within VANETs. However, standardized security mechanisms lead to significant overhead in terms of bandwidth requirement and delay. Prior work has focused on reducing the overhead by advanced strategies for pseudonym and authorization authority certificate exchange. However, we find that this is not enough to enable reliable message exchange in VANETs. Various other sources of overhead caused by security mechanisms in VANETs are identified in the provided analysis. Thereby, we find cross layer and cross message dependencies. In combination with the non-fragmentation property of VANET messages, such dependencies are discovered to lead to massive dropping of packets due to maximum size violations at low protocol layers. Thus, we develop a method for cross layer on demand content assembling for VANET messages, which can avoid the size limit violations without preventing individual layers from disseminating their variable length data sets.