Rabbit: Secure Encrypted Property Graph Search Scheme Supporting Data and Key Updates

Online social networks (OSNs) utilize property graphs to represent user data and relationships, but their reliance on public cloud infrastructures raises significant privacy and security concerns. Efficiently supporting secure and dynamic data updates in complex graphs, while ensuring the security of the newly generated ciphertexts during the key being compromised and the time when the new key is updated, is a critical challenge. In this paper, we proposed a novel dynamic encrypted property graph search scheme named Rabbit. To the best of our knowledge, Rabbit is the first cryptographic privacy-preserving graph search scheme that supports efficient multi-attribute conjunctive querying and data updates in complex graphs, with post-compromise security provided during key updates. In addition, Rabbit achieves sub-linear search efficiency, with the time complexity for both attribute-grained updates and node-grained deletions being $\mathcal{O}(1)$. Moreover, Rabbit employs a robust two-layer encryption mechanism, ensuring that even if the client’s key is compromised, the integrity and security of the newly generated ciphertexts remain unaffected.