An Overview of Secure Enclaves: Fort Knox or El Dorado?
Bio
Margot Bauman is a LEAP student in Computer Engineering, who holds graduate degrees in history and teaching. She is interested in addressing computer security issues at the hardware/architectural level.
Abstract
In the Adaptive and Secure Computing Systems Lab, we focus on hardware-level security. Therefore, when Professor Kinsy suggested “secure enclaves” as a research topic for this seminar, I approached it from the hardware side. As a historian with more than a passing interest in medieval architecture, specifically castles, the notion had additional appeal. However, that same passing interest meant I approached the topic with a significant quantity of skepticism: there are many reasons we no longer build castles and cannonballs are only one. However, computers and castles are very different animals. Therefore, I set out to determine whether the security created by trusted execution environments was, in fact, secure, or whether it was just an illusion. (Spoilers: so far, we’re not doing so well…)
References
- Costan, V. and S. Devadas. Feb. 2016. "Intel SGX Explained." Cryptology ePrint Archive, Report 2016/086.
- Costan, V., I. Lebedev, A. Wright, S. Zhang, A. Mitahl, & S. Devadas. 2016. "Sanctum Hardware Extensions for Strong Software Isolation." in 25th USENIX Security Symposium.
- Bourgeat, T., I. Lebedev, A.Wright, S. Zhang, A. Mithal, & S. Devadas. 2018. "MI6: Secure Enclaves in a Speculative Out-of-Order Processor." ArXiv, abs/1812.09822.