For over many decades, relational database model has been considered as the leading model for data storage and management. However, as the Big Data explosion has generated a large volume of data, alternative models like NoSQL and NewSQL have emerged. With the advancement of communication technology, these database systems have given the potential to change the existing architecture from centralized mechanism to distributed in nature, to deploy as cloud-based solutions. Though all these evolving technologies mostly focus on performance guarantees, it is still being a major concern how these systems can ensure the security and privacy of the information they handle. Different datastores support different types of integrated security mechanisms, however, most of the non-relational database systems have overlooked the security requirements of modern Big Data applications. This paper reviews security implementations in today's leading database models giving more emphasis on security and privacy attributes. A set of standard security mechanisms have been identified and evaluated based on different security classifications. Further, it provides a thorough review and a comprehensive analysis on maturity of security and privacy implementations in these database models along with future directions/enhancements so that data owners can decide on most appropriate datastore for their data-driven Big Data applications.

No abstract

The explosion of big data along with cloud computing architecture has empowered cloud-based database infrastructures to efficiently manage large and distributed data volumes in the cloud by facilitating numerous data-driven applications. Recently, different systems have been proposed that can additionally protect the privacy of these data by keeping them encrypted at the database level, and enabling them to make trusted query executions over encrypted data. This approach has given the potential and much safer means to outsource private information to untrusted and distributed cloud platforms. However, regardless of their proficiency towards handling a large volume of private information , these techniques are exposed to different access pattern attacks. Oblivious Random Access Machine (ORAM) is a security primitive well known for mitigating such attacks; however, direct integration of ORAM into cloud-based database systems is much more challenging due to high-performance penalties and minimal query functionalities. In this paper, we propose a novel data processing framework for database systems in the cloud using distributed ORAM techniques and oblivious data structures, making database queries resilient to access pattern attacks. We implemented our framework on a practical database setup and evaluated the performance based on different industrial metrics. The experimental results demonstrate that our distributed approach has significant benefits for cloud-based database systems compared to the direct integration of ORAM primitives at the database level.