Safe & Efficient Privacy-policy Enforcement on Hadoop

Hadoop's lack of an effective access control system raises questions about whether confidentiality and integrity can be maintained for sensitive user data in the cloud. Researchers have begun to address this need for access control, but current solutions either have unacceptable performance overhead for efficient processing, do not provide any clear understanding on levels of privacy provided or are limited in application scope. This paper proposes a lightweight access control layer on top of Hadoop for access control on shared data. Our work targets query engines built on Hadoop, which use SQL-like query languages to directly query datasets. We provide much needed fine-grained access control (FGAC), to provide a secure interface between users and target systems without noticeable overhead.