Research Mentor: John A. Chandy
Data is the lifeblood of any organization and protection of this data is vital. This research is focused on two aspects of data security. First, we investigate the use of cloud storage as a mechanism for data storage backup and the implications of privacy when using shared storage media. Particularly, when using data deduplication and bandwidth reduction techniques, can guarantees be made about the privacy and security of the data within the cloud. What do different levels of security and privacy mean in terms of performance and cost? Secondly, we evaluate an architecture that minimizes the involvement of the operating system in the dataflow of an application. Using intelligent active object storage mechanisms, the application makes direct access to data storage through a hardware abstraction layer. Application aware storage mechanisms are enabled through file specific extended attributes. The operating system is insulated from any protected data access, thus preventing sensitive data from leaking out to an insecure operating system or network.
Components for Student Participation
Student research activities include constructing an active storage infrastructure, integrating an application-aware storage platform with the active storage, providing an hardware abstraction layer, modifying parallel file systems, etc. Students will learn how to write small software test suites, build operating system kernels, and conduct benchmark evaluations. The project supervisor along with senior graduate students will work closely with the REU students and provide mentorship.