Privacy-Preserving Noisy Keyword Search in Cloud Computing

摘要

We consider the following problem: a user with either limited resources or limited expertise wants to outsource its private documens, which are associated with noisy keywords, to an untrusted cloud server in a private manner, while maintaining the ability to retrieve the stored data in a fault-tolerant manner. For example, the organization of homeland security wishes to outsource its private criminal database comprised of a set of criminal dossiers to the cloud server in encrypted form and hopes to retrieve encrypted dossiers by biometrics. In this paper, we first present a general framework for searching on private-key encrypted data by noisy keywords in a fault-tolerant manner. Then we propose a concrete scheme which is proved secure against an adaptive adversary under well-defined security definition. It achieves search in two rounds of communication, and requires an amount of work from the server that is linear in the number of noisy keywords.