Traceable Private Set Intersection in Cloud Computing

Private set intersection (PSI) is a fundamental primitive of many secure applications, which is broadly attracting researchers' interests. The most of existing private set intersection solutions are constructed based on the designed garbled Bloom filter to prevent against the semi-honest adversary. However, the security attributes can no longer be hold in the presence of the malicious adversary. In this paper, we present a novel traceable PSI protocol to protect the confidentiality of the participants in the presence of the malicious adversary, when operating on the large-scale data set. Specifically, our scheme relies on the committed Bloom filter, committed random selection and ownership proof, which enables clients and the cloud server to conduct secure and fine-grained verification through randomly checking a small fraction of clients' data items. To further reduce the communication and computation overhead, a probabilistic approach is developed to provide efficient verification. In the end, we conduct security analysis of our scheme and prove that it is secure under the full universe attack.