Towards a secure and efficient search over encrypted cloud data

Abstract

Cloud computing enables new types of services where the computational and network resources are available online through the Internet. One of the most popular services of cloud computing is data outsourcing. For reasons of cost and convenience, public as well as private organizations can now outsource their large amounts of data to the cloud and enjoy the benefits of remote storage and management. At the same time, confidentiality of remotely stored data on untrusted cloud server is a big concern. In order to reduce these concerns, sensitive data, such as, personal health records, emails, income tax and financial reports, are usually outsourced in encrypted form using well-known cryptographic techniques. Although encrypted data storage protects remote data from unauthorized access, it complicates some basic, yet essential data utilization services such as plaintext keyword search. A simple solution of downloading the data, decrypting and searching locally is clearly inefficient since storing data in the cloud is meaningless unless it can be easily searched and utilized. Thus, cloud services should enable efficient search on encrypted data to provide the benefits of a first-class cloud computing environment. This dissertation is concerned with developing novel searchable encryption techniques that allow the cloud server to perform multi-keyword ranked search as well as substring search incorporating position information. We present results that we have accomplished in this area, including a comprehensive evaluation of existing solutions and searchable encryption schemes for ranked search and substring position search.