SaTC: CORE: Small: External Obliviousness in Trusted Execution Environments

Syracuse University, Syracuse NY

Investigators

Abstract

Processing personal data in "the cloud" is an increasingly common computing application. For example, health monitors collect heartbeat data and upload to cloud for personal healthcare and wellbeing. This increases privacy and data security risks. This project aims to achieve provable security in the cloud by protecting data-access patterns in a security-oriented cloud architecture built on the recently released Intel SGX (Software Guard eXtensions) processor. The approach of the project is to combine hardware-level protection for strong processor-memory isolation with algorithm-level protection for data-oblivious computing. This new combination, termed external obliviousness, is intended to enable strongly secure computation at a large data scale. The project consists of two thrusts: 1) building externally oblivious programming support for common computation primitives, 2) building externally oblivious software systems for data-analytical queries. The project will result in multiple broader impacts: The new system-building method of external obliviousness will influence cyber-security and systems research community. The new secure software will potentially expand the cloud industry into supporting security-oriented applications in electronic healthcare. The project will lead to hands-on lab modules integrated with residential curriculum and a nation-wide education platform for both undergraduate students and under-represented groups. The project will generate performance-evaluation results from the developed codebase. They will be published to the online repository and be maintained there for 3 years: https://exo-sgx.github.io. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.