Secure Relational Data Access in The Cloud

With the success of the cloud computing paradigm, data security in the cloud remains a big concern and even an obstacle that prohibits the adoption of enterprise usage of the cloud. Although encryption can provide confidentiality for enterprise data, most current data access methods on encrypted data either sacrifice query efficiency for data confidentiality, or trade off partial confidentiality to achieve efficiency in supporting common database queries. Moreover, they do not consider data reliability, which is needed for securing data in the dynamic cloud. This paper proposes a relational data access framework in the cloud that satisfies the seemingly conflicting goals of strong data confidentiality, data reliability and query efficiency. Our solution satisfies the three goals consistently in supporting different types of queries and data operations including exact queries, range queries and data updates. Data confidentiality and reliability are achieved by using the Information Dispersal Algorithm (IDA) with a random factor to encode and disperse the data. Query efficiency is achieved by building a secure B+-tree index on the data. The index is also encoded and dispersed in the cloud using IDA. We establish the security of our proposal by a detailed security analysis argument and demonstrate the query efficiency of our proposal through an experimental evaluation.