Confidential Computing in Front-End: Enhancing Data Security with Secure Enclaves and Homomorphic Encryption

This thesis investigates how homomorphic encryption and, more especially, Secure Enclaves might improve data security in front-end web systems using Confidential Computing methods. Conventional encryption techniques are insufficient in maintaining confidentiality during data processing when client-side technologies progressively control sensitive data. This paper evaluates in real-world front-end applications the utility, efficiency, and scalability of Secure Enclaves (e.g., Intel SGX and ARM TrustZone) and numerous Homomorphic Encryption frameworks (including BFV and CKKS). Analyzing real-world events, including secure form entries and in-browser document management, helps the study assess performance indicators, including latency, CPU use, and integration complexity. Studies show that Secure Enclaves provide better performance and simpler integration; homomorphic encryption improves privacy protections even if it results in higher computing costs. This work ends with analyzing the trade-offs related to different approaches and suggesting suitable use cases for every scenario, thereby improving the safety of data while in use.