BARS: a Blockchain-based Anonymous Reputation System for Trust Management in VANETs

Zhaojun Lu, Qian Wang, Gang Qu, Zhenglin Liu

Abstract

The public key infrastructure (PKI) based authentication protocol provides the basic security services for vehicular ad-hoc networks (VANETs). However, trust and privacy are still open issues due to the unique characteristics of vehicles. It is crucial for VANETs to prevent internal vehicles from broadcasting forged messages while simultaneously protecting the privacy of each vehicle against tracking attacks. In this paper, we propose a blockchain-based anonymous reputation system (BARS) to break the linkability between real identities and public keys to preserve privacy. The certificate and revocation transparency is implemented efficiently using two blockchains. We design a trust model to improve the trustworthiness of messages relying on the reputation of the sender based on both direct historical interactions and indirect opinions about the sender. Experiments are conducted to evaluate BARS in terms of security and performance and the results show that BARS is able to establish distributed trust management, while protecting the privacy of vehicles.

The Trust-Privacy Paradox in VANETs

As Intelligent Transportation Systems (ITS) evolve, Vehicular Ad-hoc Networks (VANETs) rely on Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication to share critical safety alerts. However, these networks face a significant "trust-privacy paradox." Ensuring the security and authenticity of messages (e.g., verifying a "road hazard" alert isn't malicious) often requires linking public keys to real-world identities, which compromises driver privacy by enabling movement tracking. Existing trust models frequently suffer from data sparsity or high latency, making them impractical for high-speed traffic environments.

BARS: Architecture and Methods

To resolve this conflict, the authors propose BARS (Blockchain-based Anonymous Reputation System). The framework introduces a multi-layered approach to authentication and trust:

  • Identity-Pseudonym Separation: BARS decouples identity management from certificate issuance. A Law Enforcement Authority (LEA) maintains the link between real identities and public keys but does not issue certificates. Conversely, a Certificate Authority (CA) issues certificates without knowing the vehicle's real identity. This ensures "conditional anonymity"—privacy for the public, but accountability for the LEA in case of disputes.
  • Dual-Blockchain Authentication: The system utilizes two distinct ledgers: the CerBC (Certificate Blockchain) for verifying valid certificates and the RevBC (Revocation Blockchain) for tracking revoked keys.
  • Efficient Revocation Checks: By employing a LexTree (lexicographical Merkle tree) within the RevBC, BARS allows vehicles to provide efficient "proof of absence" for revoked keys, significantly reducing authentication time compared to traditional revocation lists.

Reputation-Based Trust Model

At the heart of BARS is a dynamic reputation system. Every vehicle maintains a reputation score that evolves based on its behavior:

  • Positive Reinforcement: Scores increase when a vehicle broadcasts honest, verified alerts. The algorithm incentivizes early reporting in sparse areas by factoring in traffic density ($D_r$) and alert urgency ($L$).
  • Penalization: Scores decrease if a vehicle is caught spreading forged messages or submitting false reports about others.
  • Message Persistence: A dedicated Message Blockchain (MesBC) records all broadcasts, providing an immutable audit trail for the LEA to update reputation scores and resolve conflicts.

Key Results and Performance

The authors evaluated BARS through numerical simulations and computational benchmarking. Key findings include:

  • Robust Security: The system successfully eliminates linkability between public keys and real identities for all parties except the LEA.
  • Operational Efficiency: Storage requirements for block headers are minimal (approx. 4.2 MB/year). In a simulation of 1 million vehicles, the transmission overhead for anonymous authentication was roughly 0.177 M/s, suitable for real-world deployment.
  • Behavioral Dynamics: Simulations demonstrate that the reputation algorithm effectively distinguishes between honest and malicious actors, as shown in the reputation score fluctuations over time.

Limitations and Future Directions

While BARS offers a robust solution, several challenges remain:

  • Consensus Latency: The reliance on Proof of Work (PoW) at Roadside Units (RSUs) may introduce latency in ultra-fast traffic. Future iterations could explore Proof of Stake (PoS) to reduce energy consumption and improve speed.
  • Centralization Risks: The LEA remains a single point of failure for privacy. If compromised, the anonymity of the entire network is at risk.
  • Network Volatility: Further testing is needed to assess performance in "shadow zones" with intermittent RSU connectivity or extreme network volatility.

Summary updated

BibTeX
@inproceedings{lu2018bars,
  author = {Zhaojun Lu and Qian Wang and Gang Qu and Zhenglin Liu},
  title = {{BARS: a Blockchain-based Anonymous Reputation System for Trust Management in VANETs}},
  booktitle = {TrustCom/BigDataSE},
  year = {2018},
  doi = {10.1109/trustcom/bigdatase.2018.00025},
}
Figure