Paper 2023/1631

ASKPIR: Authorized Symmetric Keyword Privacy Information Retrieval Protocol Based on DID

Abstract

Symmetric Private Information Retrieval (SPIR) is a stronger PIR protocol that ensures both client and server privacy. In many cases, the client needs authorization from the data subject before querying data. However, this also means that the server can learn the identity of the data subject. To solve such problems, we propose a new SPIR primitive, called authorized symmetric keyword information retrieval protocol (ASKPIR). Specifically, we designed an efficient DID identification algorithm based on the Pedersen Commitment, which is used to solve the identity management and privacy problems of data subject when data is shared by multiple parties in a distributed environment. Then, we present a novel authorization algorithm combining NIZK proof and DID, which can preserve client privacy. Finally, to improve the efficiency of client retrieval, our protocol constructs PSI-Payload with mqRPMT and OTE so as to support batch keyword searches. In addition, we provide a formal security analysis for the anonymity and unforgeability of the protocol and demonstrate that ASKPIR can achieve malicious security under the UC framework. Theoretical analysis and experimental results show that the ASKPIR protocol is more efficient than other related works and solves the problem of incompatibility between data subject authorization and client privacy.