Data intensive scalable infrastructure for data-driven research: Big data and warehouse architecture

Summary form only given, as follows. The complete presentation was not made available for publication as part of the conference proceedings. This paper outlines the design and develop, a new type of data analytics infrastructure for secure and privacy-preserving data analytics designed to meet the needs of data-driven research. The infrastructure ecosystem allow data owners to attach security and privacy policies to their data while the system ensures that the policies are enforced. The ecosystem provides strong, provable, end-to-end guarantees, for researchers to use, and approach the performance of today's data analytics frameworks particularly in terms of developing new HPC workflows. The Data Intensive Scalable Infrastructure for Data-Driven Research (DISI-DDR) ecosystem improves on existing data analytics infrastructure by admitting policies that limit what the analytics program may do with the data a type of limitation not possible in systems based on access control. This is achieved with the just-released Intel Software Guard Extensions (SGX) trusted hardware platform to address the second. The resulting infrastructure offers end-to-end security and privacy guarantees without requiring analysts to write security checks, andl offer similar performance to existing insecure analytics frameworks. The ecosystem is designed to eliminate the duplicated effort common to data-driven research. Researchers in these fields are commonly forced to build new infrastructure to support data protection and secure collaboration; despite this effort, most such systems provide weak security and privacy guarantees. Also, the ecosystem allows the secure collection of data from smartphone users and privacy-preserving analysis of that data.