Outsourcing of Verifiable Attribute-Based Keyword Search

Abstract

In integrated broadcast-broadband services, viewers receive content via the airwaves as well as additional content via the Internet. The additional content can be personalized by using the viewing histories of each viewer. Viewing histories however contain private data that must be handled with care. A verifiable attribute-based keyword search (VABKS) scheme allows data users (service providers), whose attributes satisfy a policy that is specified by the data owner (viewer), to securely search and access stored data in a malicious cloud server, and verify the correctness of the operations by the cloud server. VABKS, however, requires data owners who have computationally weak terminals, such as television sets, to perform heavy computations due to the attribute-based encryption process. In this paper, we propose a new VABKS scheme where such heavy computations are outsourced to a cloud server and hence the data owner is kept as light as possible. Our scheme is provably secure against two malicious cloud servers in the random oracle model: one performing the attribute-based encryption process, and the other performing the keyword search process on the encrypted data. We implement our scheme and the previous VABKS scheme and show that our scheme significantly reduces the computation cost of the data owner.