13-09-2017, 12:59 PM
Cloud computing economically allows the paradigm of outsourcing data services. However, in order to protect data privacy, sensitive cloud data must be encrypted before being outsourced to the commercial public cloud, which makes the data usage service a very difficult task. Although traditional search encryption techniques allow users to securely search encrypted data through keywords, they only support Boolean search and are not sufficient to meet the need for effective data usage that is inherently demanded by a large number of users and large amount of data files in the cloud. In this article, we define and solve the problem of finding secure classified keywords on encrypted data in the cloud. The classified search greatly improves the usability of the system by allowing the ranking of relevance of search results instead of sending undifferentiated results and also ensures the accuracy of file retrieval. Specifically, we explore the approach of statistical measurement, ie, relevance score, retrieval of information to construct a secure search index, and develop a one-to-many order preservation mapping technique to adequately protect such information sensitive score. An in-depth analysis shows that our proposed solution enjoys a security guarantee "as strong as possible" compared to previous encryption schemes, while also correctly accomplishing the goal of classified keyword search.