PINQ: Privacy-Preserving Influential Nodes Query Over Encrypted Distributed Networks

Identifying influential nodes in complex networks is crucial for information dissemination, epidemic control, and supply chain management, etc. As the scale of networks increases, data owners prefer to outsource identifying influential nodes to cloud servers. To protect data privacy, network data needs to be encrypted before being outsourced. So identifying influential nodes on encrypted data is necessary. Furthermore, many sensitive networks are distributed among different data owners. To identify the influential nodes in such networks, it is necessary to first obtain the whole network and then outsource the whole network to supply identify influential nodes service. Currently, there is only one scheme that supports privacy-preserving identification of influential nodes in distributed private networks, but it works on a wrong whole network, it leaks some topological characteristics of the whole network, can only support influential nodes identifying, and cannot support private querying for any influential nodes. To address these issues, we propose PINQ (Privacy-preserving Any Influential Node Query Scheme), the first general scheme that supports any influential node queries in private distributed networks. PINQ can query the top-k most influential nodes, the bottom k influential nodes, or nodes with influence at a specific rank. PINQ uses encryption-selection-re-randomization techniques to correctly generate the privacy whole network within distributed private networks, securely ranking the importance of nodes without leaking any network topology information, and supporting any influential node privacy queries. Additionally, PINQ uses blockchain to achieve fair data query transactions. Theoretical analysis and experimental simulation results show that PINQ is secure, feasible and efficient.