Homomorphic Encryption Task Segmentation for TEE Security and Performance

The patent segments the homomorphic computing task into two parts: security-sensitive operations are executed outside TEE using pure software homomorphic encryption, while non-security-critical operations are executed inside TEE for performance acceleration. This segmentation allows the system to simultaneously achieve both security and performance by assigning different operations to different execution environments based on their security requirements.

The patent divides the computing workload into security-critical and non-security-critical components. Non-security-critical operations are offloaded to TEE hardware acceleration, while security-critical operations remain in pure software. This segmentation enables the system to achieve high performance without compromising security by strategically placing different operations in appropriate execution environments.

The patent applies local quality by assigning different execution qualities to different parts of the computing task. Security-critical operations receive the highest security quality (executed outside TEE), while non-security-critical operations receive performance-optimized quality (executed inside TEE). This localized quality assignment allows the system to optimize overall performance while maintaining maximum security where required.

The patent extracts security-critical operations from the TEE environment and executes them in pure software homomorphic encryption. This extraction removes the vulnerability to side channel attacks from the security-sensitive parts of the system while retaining TEE acceleration for non-security-critical operations, thereby reducing overall dependency on hardware manufacturers and mitigating side channel attack risks.