Efficient and Secure Multi-Party Private Set Operation Protocol Suitable for IoT Devices
Yichen Bao, Lifei Wei, Dezhi Han · 2023
In the IoT environment, a large amount of sensitive data are transmitted and processed by IoT devices, such as personal identity, location data and health data. To protect these sensitive data from leakage, multi-party private set operation (MPSO) has become one of the important multi-party secure computing techniques in the IoT domain, which allows the parties with private sets to perform secure intersection and union operations respectively without disclosing private information other than the results. It allows the exchange and share of IoT data effectively. Most of them use homomorphic encryption with expensive computational overhead or oblivious transfer with expensive communication costs, which can affect the performance of practical applications, especially for the IoT devices. A new multi-party multi-query set membership testing component is designed to allow multiple participants to compute the differences of their own datasets. In addition, a multi-party private set operation protocol is constructed based on this component to achieve minimal overhead for communication in a star network topology, and to designate the central party to obtain the final computation results. The intersection or union results is computed while obtaining the cardinality of the set intersection without revealing any additional information. The security of the constructed protocol is proved under a semi-honest model. The proposed framework is an efficient extension of the existing work mqRPMT. The experiment is using the elliptic curve curve25519 on ECDH key negotiation. In the multi-party private set intersection protocol, the proposed protocol has fewer communication rounds and higher computational efficiency than those of the existing work for small sets. In the multi-party private set union protocol, the proposed protocol runs in O(nlogn) time complexity and constant communication rounds, which is better than the existing work.