EPRV: Efficient Pseudonym Revocation in VANETs
Noureddine Chaib, Nasreddine Lagraa, Mohamed Bachir Yagoubi, Mohamed Lahcen Bensaad and Abderrahman Lakas
Vehicular ad hoc networks are vulnerable to insider attacks like dropping safety messages and injecting bogus ones. A typical solution to this problem is to use the local revocation by making each vehicle responsible for monitoring its neighbours and notifying the newcomers through accusation messages. There are multiple solutions that have been proposed for this purpose. But, the impact of pseudonym change on these solutions has not been well studied. Moreover, the proposed solutions consist in checking the revocation condition about a vehicle only upon receiving an accusation message against it. In contrast, the revocation condition might hold whenever the accusation list content changes. This work is devoted to study the impact of pseudonym change on revocation approaches. Particularly, we are interested in studying the impact of malicious vehicles that do not respect the pseudonym change policy. In addition to that, these vehicles amplify the number of falsified accusation messages through a quick change of pseudonyms. To that effect, we have defined a new concept that we called “Accusation community” to model revocation schemes. We have also proposed two new mechanisms DAPM (Duplicate Accusations Prevention Mechanism) and IRM (Instantaneous Revocation Mechanism) to ensure the correctness of revocation schemes and to enhance their performance. We have implemented these two mechanisms in a new revocation solution that we called EPRV (Efficient Pseudonym Revocation in VANETs). We have conducted multiple simulation scenarios to analyze the performance of EPRV. The obtained results show that it largely boosts the revocation performance in function of false positives, detection rates and the involved latency.
Keywords: VANET, Revocation, Pseudonym, Malicious, Detection