Comparative Performance Evaluation of BFV, BGV and CKKS Homomorphic Encryption Schemes Using Microsoft SEAL

Abstract

Homomorphic Encryption (HE) enables arithmetic operations to be performed directly on encrypted data, thereby preserving confidentiality during computation. However, the high computational cost associated with HE continues to limit its large-scale practical deployment. This paper presents a comparative experimental evaluation of three prominent homomorphic encryption schemes, Brakerski/Fan-Vercauteren (BFV), Brakerski-Gentry-Vaikuntanathan (BGV), and Cheon-Kim-Kim-Song (CKKS), using the Microsoft SEAL library. The study investigates the performance of fundamental homomorphic operations, including encryption, decryption, addition, multiplication, and squaring, across multiple polynomial modulus degree settings. Experiments were conducted using Microsoft SEAL’s benchmarking framework, and performance outputs were systematically analyzed to identify runtime trends and operational trade-offs. The results show that execution time increases substantially with higher parameter sizes for all three schemes, reflecting the inherent balance between security strength and computational efficiency. While ciphertext addition incurred minimal overhead, multiplication and squaring represented the dominant performance bottlenecks. Furthermore, CKKS showed relatively competitive efficiency in selected computational scenarios compared to BFV and BGV. The findings contribute empirical insight into the practical behavior of widely used HE schemes and provide guidance for selecting appropriate schemes and parameters in privacy-preserving computing environments.