Impact of endorsement policy on the performance of blockchain-based VANET

With the revolution in technology, VANET has emerged as a new generation of technology and used globally for intelligent traffic management systems. VANET uses a PKI-based protocol to maintain privacy and security. Still, some crucial issues remain unresolved, like blind belief in communicated messages, lack of vehicle's interest in participation, revealing the sender vehicle's identity, and maintaining integrity. The solution to these issues will prevent vehicles from propagating fake messages while concurrently protecting the confidentiality of each involved vehicle against tracking attacks. Blockchain, a peer-to-peer, immutable, distributed ledger-based technology, can be adapted as a robust solution to overcome critical challenges and make VANET more secure and reliable. This paper proposes a novel blockchain-based VANET architecture, blockchain-based vehicular ad-hoc network (B-VANET), implemented on the Hyperledger Fabric platform. B-VANET aims to secure the network by validating road events before actually updating it on the network. For validation, the trust factor of each vehicle and their past involvement in the network is maintained. Trust factor value decides the trustworthiness of a vehicle. Based on this trust factor, a dynamic endorsement policy is designed. Endorsement policy, part of the smart contract, plays a vibrant role in Hyperledger Fabric. It decides how many and which particular peers can endorse the transaction to make it valid and add it to the ledger. Multiple possible approaches to endorsement policies are identified for B-VANET and are analyzed using the Hyperledger Caliper tool. These approaches are evaluated for data integrity, confidentiality, and real-time system issues. An appropriate approach for the B-VANET system is recognized, giving the best throughput and latency in all possible scenarios. It is analyzed that compared to the worst approach, the best approach achieves an almost 15% increase in throughput and reduces latency by nearly half. We also compared the performance of B-VANET and other approaches to show that our proposed system better meets VANET's security needs.