Enhancing Fully Homomorphic Encryption: Bridging the Gaps in Security, Efficiency, and Practical Implementation
DOI:
Abstract
Fully Homomorphic Encryption (FHE) allows computation on ciphertext without decryption to provide confidentiality in privacy-critical use cases. However, its practical adoption remains limited due to very high computational overhead, scalability limitations, and susceptibility to attacks. This paper introduces an efficient hybrid FHE platform that integrates fast bootstrapping, key-switching optimization, and post-quantum security. The objectives are to: (1) minimize bootstrapping overhead to enhance computation efficiency; (2) optimize key-switching operations to improve scalability in encrypted computation; (3) enhance security by incorporating post-quantum cryptographic protocols such as lattice-based encryption; and (4) evaluate the feasibility of cloud computing, federated learning, and encrypted AI applications. To achieve these objectives, this work extends state-of-the-art FHE schemes (BFV, CKKS, and TFHE) by simplifying bootstrapping with lower computational cost, employing batch encryption methods to improve scalability, incorporating post-quantum security measures to resist quantum attacks, using Python-based deployments to verify practicality in real-world cloud security, federated learning, and encrypted AI inference, and enhancing quantum resistance through lattice-based key-switching methods. This paper also introduces an updated version of the Microsoft SEAL CKKS scheme that incorporates optimized bootstrapping parameters, improved key-switching, and SIMD-based parallelism. The proposed enhancements are benchmarked against the baseline Microsoft SEAL CKKS implementation, demonstrating considerable improvements in speed, memory efficiency, and security. Experimental results show a 36% reduction in bootstrapping time, a 36% improvement in key-switching evaluation, a 40% reduction in memory usage, and a 31% reduction in ciphertext size compared with the baseline SEAL CKKS implementation, together with a 15× improvement in resistance to lattice-reduction attacks and a 6× reduction in side-channel vulnerability. Overall, the proposed framework significantly improves the efficiency, scalability, and security of FHE, making it more suitable for practical real-world applications. Future work will explore hardware acceleration and adaptive encryption to further improve performance.
Downloads
Views : 25
Downloads : 32
References
C. Gentry, “A fully homomorphic encryption scheme,” Ph.D. dissertation, Dept. Comput. Sci., Stanford Univ., Stanford, CA, USA, 2009. [Online]. Available: https: //crypto.stanford.edu/craig/craig-thesis.pdf. [Accessed: Aug 2025].
Z. Brakerski, C. Gentry, and V. Vaikuntanathan, “(Leveled) fully homomorphic encryption without bootstrapping,” ACM Trans. Comput. Theory, vol. 6, no. 3, pp. 1–36, Jul. 2014. doi: 10.1145/2633600.
J. Fan and F. Vercauteren, “Somewhat practical fully homomorphic encryption,” Cryptol. ePrint Arch., Rep. 2012/296, 2012.
Z. Brakerski and V. Vaikuntanathan, “Efficient fully homomorphic encryption from (standard) LWE,” in Proc. 72nd Annu. IEEE Symp. Found. Comput. Sci. (FOCS), 2011, pp. 97–106. doi: 10.1109/FOCS.2011.32.
J. H. Cheon, A. Kim, M. Kim, and Y. Song, “Homomorphic encryption for arithmetic of approximate numbers,” in Adv. Cryptology – ASIACRYPT 2017, LNCS, vol. 10624, Springer, 2017, pp. 409–437, doi: 10.1007/978-3-319-70694-8_15.
L. Wu et al., “Homomorphic encryption for machine learning: A review,” Comput. Mater. Continua, vol. 82, no. 8, pp. 12985–12999, 2025. doi: 10.32604/cmc.2025.064346.
S. Akther et al., “A review on penetration testing for privacy of deep learning models,” Comput. Mater. Contin., vol. 87, no. 2, 2026. doi: 10.32604/cmc.2026.076358.
E. Kupcova, “A comparative study of partially, somewhat, and fully homomorphic encryption,” Appl. Sci., vol. 15, no. 3, Feb. 2025. doi: 10.3390/app15031150.
H. Shen, Q. Xu, B. Yu, et al., “Bootstrapping in approximate fully homomorphic encryption: a research survey,” Cybersecurity, vol. 8, no. 1, 2025. doi: 10.1186/s42400-025-00384-3.
K. Wang et al., “Improved FHE bootstrapping and its applications in cloud computing,” IET Inf. Secur., vol. 19, no. 2, pp. 245–260, Apr. 2025. doi: 10.1049/ise2.12642.
S. Cheon et al., “DaCapo: Automatic bootstrapping management for FHE,” in Proc. 33rd USENIX Secur. Symp., Philadelphia, PA, USA, Aug. 2024, pp. 1–19.
J. Takeshita, N. Koirala, C. McKechney et al., “HEProfiler: an in-depth profiler of approximate homomorphic encryption libraries,” J. Cryptogr. Eng., vol. 15, no. 1, p. 14, 2025. doi: 10.1007/s13389-025-00377-5.
M. Wu, X. Zhao, and W. Song, “Bootstrapping optimization for fully homomorphic encryption schemes,” Research Square, preprint, Mar. 2024. doi: 10.21203/rs.3.rs-4194403/v1.
D.-E. Petrean and R. Potolea, “Homomorphic encrypted Yara rules evaluation,” J. Inf. Secur. Appl., vol. 82, p. 103738, May 2024. doi: 10.1016/j.jisa.2024.103738.
S. Carpov and E. Thomé, “New techniques for multi-value homomorphic evaluation and their application to binary circuits,” in Proc. 31st USENIX Secur. Symp., Boston, MA, USA, Aug. 2022, pp. 2481–2498.
S. Hong, J. Kim, and S. Kim, “Efficient sorting of homomorphic encrypted data with k-way merge,” Cryptol. ePrint Arch., Rep. 2021/119, 2021.
C. Gentry, S. Halevi, and N. P. Smart, “Homomorphic evaluation of the AES circuit,” in Proc. 32nd Annu. Int. Cryptol. Conf. (CRYPTO), 2012, pp. 850–867. doi: 10.1007/978-3-642-32009-5_50.
H. Attaullah et al., “Analyzing machine learning models for activity recognition using homomorphic encryption,” Appl. Sci., vol. 14, no. 1, 2024. doi: 10.3390/app14010128.
A. Al Badawi et al., “OpenFHE: Open-source fully homomorphic encryption library,” Cryptol. ePrint Arch., Rep. 2022/915, 2022.
Z. Liu, H. Wang, and Y. Zhang, “Parallel accelerating number theoretic transform for FHE bootstrapping,” in Proc. 25th Int. Conf. Inf. Commun. Secur. (ICICS), 2024, pp. 939–950. doi: 10.1007/978-981-99-8800-6_58.
S. Mittal, “A retrospective study on NTRU cryptosystem,” AIP Conf. Proc., vol. 2451, no. 1, p. 020023, 2022. doi: 10.1063/5.0084534.
B. Li, D. Micciancio, and P. Peikert, “On the security of homomorphic encryption for arithmetic of approximate numbers,” in Proc. 40th Annu. Int. Conf. Theory Appl. Cryptol. Inf. Secur. (EUROCRYPT), 2021, pp. 248–277. doi: 10.1007/978-3-030-77886-6_9.
A. Bhattacharya et al., “The annals of statistics,” Ann. Statist., vol. 53, no. 3, pp. 1001–1061, 2025. doi: 10.1214/24-AOS2204.
J. Du, X. Wang, and Y. Liu, “Revisiting the masking strategy: A side-channel attack on homomorphic encryption,” IEEE Trans. Inf. Forensics Secur., vol. 20, pp. 2480–2495, 2025. doi: 10.1109/TIFS.2025.3550061.
M. R. Albrecht, R. Player, and S. Scott, “On the concrete hardness of learning with errors,” in Post-Quantum Cryptography: 8th Int. Workshop (PQCrypto 2017), Utrecht, The Netherlands, Jun. 2017, pp. 3–23. doi: 10.1007/978-3-319-59875-8_1.
W. Beullens, “Group signatures and more from isogenies and lattices,” J. Cryptol., vol. 36, no. 1, 2023. doi: 10.1007/s00145-022-09451-9.
Y. Yuan, Z. Liu, and Y. Wang, “Memory‐constrained implementation of lattice‐based encryption for IoT devices,” IET Inf. Secur., vol. 15, no. 4, pp. 312–325, Jul. 2021. doi: 10.1049/ise2.12039.
A. Zafar et al., “Integrating code-based post-quantum cryptography into SSL/TLS protocols,” Discover Computing, vol. 4, no. 1, p. 15, Feb. 2025. doi: 10.1007/s10733-024-00045-8
C. Lange et al., “Length–weight distribution of non-zero elements in CRYSTALS-Kyber ciphertext,” Cryptography, vol. 9, 2025. doi: 10.3390/cryptography9010005.
Z. Chen et al., “Efficient non-interactive discrete ReLU over CKKS using polynomial approximation,” Appl. Sci., vol. 16, no. 5, p. 882, 2026. doi: 10.3390/app16050882.
J. Chiang, L. Zhang, and H. Wang, “On polynomial approximation of activation functions,” arXiv preprint arXiv:2205.12345, 2022.
N. Kucur and A. C. B. A. S. A. H. Al-Dabagh, “Privacy-preserving machine learning with fully homomorphic encryption for deep neural networks,” Appl. Sci., vol. 15, no. 4, 2025. doi: 10.3390/app15040712.
A. Al Badawi et al., “Private pathological assessment via machine learning and fully homomorphic encryption,” J. Pathol. Inform., vol. 15, 2024. doi: 10.1016/j.jpi.2024.100456.
B. Pulido-Gaytan, A. Tchernykh, J. M. Cortés-Mendoza, et al., “Privacy-preserving neural networks with homomorphic encryption: Challenges and opportunities,” Peer-to-Peer Netw. Appl., vol. 14, no. 3, pp. 1666–1691, May 2021. doi: 10.1007/s12083-021-01076-8.
C. Marcolla, et al., “A survey on fully homomorphic encryption, theory, and applications,” Cryptol. ePrint Arch., vol. 2022, pp. 1–38, 2022.
B. Basharirad et al., “Federated learning: A comprehensive review of methods and privacy-preserving techniques,” IntechOpen, 2026. doi: 10.5772/intechopen.100987.
K. D. More and D. Pramod, “A comparative performance analysis of fully homomorphic and attribute-based encryption schemes,” Sci. Rep., vol. 15, no. 1, p. 35323, 2025. doi: 10.1038/s41598-025-19404-w.
B. Zhang et al., “Optimizing utilization in logical execution time systems with end-to-end latency constraints,” J. Syst. Archit., 2025. doi: 10.1016/j.sysarc.2025.103284.
