Efficient Updateable Private Set Intersection on Outsourced Datasets

Private Set Intersection (PSI) is a widely used cryptographic protocol in real-life applications. Nowadays, delegated PSI protocol is also being used in the cloud computing field. Traditional PSI protocols are oriented to local datasets and run with clients interacting. However, delegated PSI protocol allows parties to outsource the storage of private datasets while having the ability to delegate PSI computation to cloud servers securely. Two significant shortcomings of existing delegated PSI protocols are the lack of efficient support for updating outsourced datasets and multiple client PSI. This paper presents an updated delegated PSI protocol that can be easily extended to multiple clients. Clients can upload their datasets after encrypting them and then delegate the entire process of intersection computation to a cloud server. Our protocol also enables efficient updating of outsourced data. We use an Elastic Bloom Filter to store data and compute intersections, which can dramatically improve intersection computation efficiency. Theoretically, we analyze the complexity of the protocol and prove its security in a semi-honest model. We compare it with related protocols and demonstrate its superiority in computational performance.