Privacy Preserving Search Techniques over Encrypted Outsourced Data

Abstract

During the last decade, various type of cloud services have encouraged individuals and enterprises to store personal data in the cloud. Despite its flexibility, cost efficiency, and convenient service, protecting security and privacy of the outsourced data has always been a primary challenge. Although data encryption retains the outsourced data's security and privacy to some extent, it does not permit traditional plaintext keyword search mechanisms, and it comes at the cost of efficiency. Hence, proposing an efficient encrypted cloud data search service would be an important step forward. To address this challenge scholars introduced Searchable Symmetric Encryption (SSE) in which a client is able to perform searches over encrypted documents. However, these schemes suffer from private information leakage such as access pattern, search pattern, and co-occurrence leakage. Several recent papers show that how this critical information can be exploited to collapse the whole security system and an adaptive attacker can reveal plaintext data. In this thesis, we address the challenging problems of access pattern, search pattern, and co-occurrence private information leakage. We design and construct three schemes to tackle the above challenges. We formally prove that all of our schemes are secure and achieve a higher level of privacy by preventing and obfuscating the private information leakages. Moreover, our performance analyses demonstrate the practicality and efficiency of our approaches.