Efficient Threshold and Unbalanced Quorum PSI Protocols for Secure Data Sharing in Cloud-Assisted Internet of Things
Zhenhua Liu, Jiang Deng, Han Yu, Baocang Wang · IEEE Internet of Things Journal · 2025
Threshold private set intersection (t-PSI) protocol enables the secure determination of whether the intersection between two sets meets a specified threshold, facilitating privacy-preserving data collaboration in cloud computing for internet of things (IoT). However, existing t-PSI protocols face challenges in efficiency and scalability. To tackle these issues, we present novel t-PSI protocols for balanced and unbalanced settings, and further develop an augmented version—unbalanced quorum PSI (q-uPSI) protocol. For balanced settings, oblivious key-value store (OKVS) and additively homomorphic encryption (AHE) are used as foundational building blocks, with OKVS decoding and additively homomorphic operations replacing the costly fully homomorphic computations in previous t-PSI protocols, thus achieving linear communication and computation complexities. In unbalanced settings, we achieve sublinear communication and computation complexities with respect to the larger set by replacing certain fully homomorphic operations in the previous protocols with oblivious key-value retrieval (OKVR) and AHE. According to the experimental results, compared to the most advanced protocols available, our balanced t-PSI protocol demonstrates computation reduction from 83 95.8%. The unbalanced t-PSI protocol significantly reduces communication overhead while also improving computation efficiency. For the q-uPSI protocol, taking the number of servers is 10 as an example, we observe substantial communication improvements across all experimental parameters, with at least 83.13% reduction. Additionally, for specific parameters, the q-uPSI protocol achieves computation reduction ranging from 7.2 69.79%. Finally, we validate the security and efficiency of the proposed protocols in cloud IoT scenarios involving interest matching and medical record analysis. All protocols ensure security under the semi-honest adversarial model and provide more efficient solutions for secure IoT data sharing and collaboration in cloud environments.