Paper 2022/738 Original: eprint.iacr.org/2022/738 ↗

Secure Search on Multi-key Homomorphically Encrypted Data with Finite Fields

Abstract

Homomorphic Encryption (HE) is a very attractive solution to ensure privacy when outsourcing confidential data to the cloud, as it enables computation on the data without decryption. As the next step, searching this homomorphic data becomes necessary to navigate it in the server. In this paper, we propose a novel algorithm to search homomorphically encrypted data outsourced to an untrusted server and shared with multiple users. We optimize the steps involved in the process to reduce the number of rounds of communication. We use an order-preserving encoding to batch the data with multi-key HE cryptosystems to reduce the multiplicative depth of the equality circuits and enable direct comparison. Further, we use LEAF to retrieve indices securely, and SealPIR to retrieve the values obliviously to the user. Overall, we provide an efficient end-to-end framework for searching shared data in a semi-honest server.