Privacy-preserving reachability query services

Due to the massive volume of graph data from a wide range of recent applications and resources required to process numerous queries at large scale, it is becoming economically appealing to outsource graph data to a third-party service provider (), to provide query services. However, cannot always be trusted. Hence, data owners and query clients may prefer not to expose their data graphs and queries. This paper studies privacy-preserving query services for a fundamental query for graphs namely the reachability query where both clients' queries and the structural information of the owner's data are protected. We propose privacy-preserving 2-hop labeling (pp-2-hop) where the queries are computed in an encrypted domain and the input and output sizes of any queries are indistinguishable. We analyze the security of pp-2-hop with respect to ciphertext only and size based attacks. We verify the performance of pp-2-hop with an experimental study on both synthetic and real-world datasets.