Ridra: A Rigorous Decentralized Randomized Authentication in VANETs

Abstract

Ensuring the security and privacy of vehicle is one of the critical requirements for the safety and reliability of vehicular ad hoc networks. A variety of (conditional) anonymous authentication schemes, including group/ring signatures, pseudo-identity-based and PKI-based approaches, have been proposed to achieve highly effective privacy-preserving authentications. A recent effort, i.e., randomized authentication, leverages homomorphic encryption for vehicles to self-generate authenticated identities to achieve full anonymity. Notwithstanding a very attractive feature to prevent single-party traceability, randomized authentication faces a great challenge on the centralized data updating and the frequent clock synchronizations. It also fails to meet the necessity of non-repudiation. In this paper, we present a rigorous decentralized randomized authentication framework with conditional privacy preservation. We use homomorphic encryption and a one-way hash chain for a vehicle to self-generate randomized pseudo-identities. We deploy the pseudonym validation mechanism over the roadside units in order to support decentralized mutual identity authentication and ownership validation of vehicles in a loosely coupled or a compound manner. Our framework can provide rigorous Level 3 privacy and traceability of vehicles. We also provide a security condition on valid random values to ensure the uniqueness of pseudonym and non-repudiation of vehicles. The performance evaluation shows that our framework is generally more efficient on infrastructures in terms of computational and communication overheads than the state-of-the-art randomized authentications.