Quantitative and Qualitative Investigations into Trusted Execution Environments

I analyzed both Homomorphic Encryption (HE) and Trusted Execution Environments (TEEs) in an effort to develop a quantitative and qualitative framework to assist computer science researchers and software engineers with determining whether they should integrate a TEE into their system design. The results of this investigation indicate that under normal conditions, TEEs offer performance that can be an order of magnitude faster than HE for general computational tasks. This suggests that TEEs are much more viable for performance intensive applications than was previously thought. However, when utilizing a library OS to simplify application development for TEE platforms, TEE performance degrades, and HE techniques can outperform TEEs in unexpected scenarios. This suggests that developing frameworks for choosing whether TEEs or HE is more suitable for privacy-preserving computation is more difficult than previously thought. This dissertation provides qualitative suggestions, based on the current state-of-the-art functionalities of leading TEEs and the security/privacy expectations of developers.