Secure Computation in Untrusted Environments via Trusted Execution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for providing provable privacy guarantees for machine learning and analytics methods in an efficient manner, especially in untrusted computing environments where sensitive information is processed in rich models like deep learning models.
Innovation Solution
A computer-implemented method and system that obtains encrypted data from a storage location in an untrusted environment, performs secure computations on the encrypted data, and provides results of these computations while maintaining data privacy, using techniques such as homomorphic encryption and secure enclaves to ensure differential privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is processed in untrusted computing environments, then computational efficiency and model performance are improved, but data privacy and security guarantees deteriorate
Solution Approach 1:
The patent introduces trusted execution environments (TEEs) as intermediary components between the untrusted computing environment and the data. These TEEs act as mediators that provide cryptographic guarantees for data privacy while enabling efficient computation. The TEEs include secure enclaves that isolate sensitive operations and prevent unauthorized access, thus resolving the contradiction between computational efficiency in untrusted environments and data privacy guarantees.
2Reliability
If encryption is applied to data throughout computation, then data privacy is improved, but computational complexity and processing overhead increase
Solution Approach 1:
The patent segments the computation process into distinct phases: encrypted data processing within trusted execution environments, and result decryption only when needed. This segmentation allows standard encryption algorithms to be used for data protection while minimizing the computational overhead. The computation is divided such that sensitive operations occur within TEEs that provide hardware-level security, reducing the overall computational complexity compared to fully homomorphic encryption approaches.
Solution Approach 2:
The patent changes the operational parameters of the computing environment by introducing trusted execution environments with specific security properties. These TEEs provide cryptographic guarantees through hardware-based trust roots and secure key management, allowing efficient encrypted computation without the full overhead of software-based homomorphic encryption. The parameter change from general-purpose computing to TEE-accelerated computing resolves the complexity-privacy tradeoff.
Data Source
AI summary
A computer-implemented method according to one embodiment includes obtaining, at an untrusted environment, encrypted data from a storage location, initiating, within the untrusted environment, a performance of one or more secure computations on the encrypted data, and providing, within the untrusted environment, results of performing the one or more secure computations on the encrypted data.