Poster: Revocation in VANETs Based on k-ary Huffman Trees

TL;DR

This paper proposes an efficient revocation scheme for VANETs using k-ary Huffman trees and duplex SHA-3 hashes to optimize insertion and search operations, reducing proof size for frequently queried revoked certificates.

Contribution

It introduces a novel revocation structure combining k-ary Huffman trees with duplex SHA-3 hashing to improve efficiency in VANET certificate revocation management.

Findings

Optimized insertion and search in revocation trees.

Reduced revocation proof size for high-query vehicles.

Method to calculate optimal k for k-ary trees.

Abstract

One of the biggest problems of vehicular ad-hoc networks is revocation. The efficient management of such issue has become one of the major paradigms in this area of research. A solution proposed here is based on the use of authenticated data structures like revocation trees to replace the classical and inefficient certificate revocation lists. In particular, the idea of this paper is to propose the use of k-ary hash trees, Huffman coding and a duplex version of the SHA-3 hash function, to optimize insertions and searches in the revocation structure. Thus, the inclusion of a new certificate revoked in the tree, only implies a new iteration of the duplex construction of the hash function, avoiding recalculating the entire hashes and the entire tree. Furthermore, a k-ary Huffman tree is used to insert leaf nodes at different levels so that those revoked nodes that are more queried, are located closer to the root node position, so the revocation proof is smaller for those vehicles that spend more time on the roads. This paper details a method to calculate the optimum value $k$ for the k-ary tree in order to optimize the revocation proof size. Therefore, the proposal described here improves both the insertion of new revoked certificates in the revocation structure and the search of revoked certificates in the revocation structure. This paper is part of a work in progress, so that we plan to implement the scheme in real scenarios to get ideal values of the parameters and comparisons with other schemes.