Cross-Platform Benchmarking of the FHE Libraries: Novel Insights into SEAL and OpenFHE
Faneela, Jawad Ahmad, Baraq Ghaleb, Sana Ullah Jan, William J. Buchanan · 2025
Read Cross-Platform Benchmarking of the FHE Libraries: Novel Insights into SEAL and OpenFHE by Faneela, Jawad Ahmad, Baraq Ghaleb, Sana Ullah Jan, William J. Buchanan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-11216.txt.
Introduction
Faneela and co-authors present a comparison from two thousand twenty-five of two software libraries for computing on encrypted data. This method, called fully homomorphic encryption, lets a cloud server process data without ever seeing it in plain form. Privacy needs make the method attractive, but it is heavy and slow, so library choice matters.
The team benchmarked the SEAL and OpenFHE libraries across Linux and Windows. They measured speed, memory use, and scaling for two popular encryption schemes. Linux ran the computations faster than Windows in their tests. OpenFHE came out as the stronger choice across diverse settings. The paper aims to guide researchers and builders who need private computation in real systems.
Catalog record
| CREATOR | Faneela, Jawad Ahmad, Baraq Ghaleb, Sana Ullah Jan, William J. Buchanan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-11216
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Cross-Platform Benchmarking of the FHE Libraries: Novel Insights into SEAL and OpenFHE Faneela, Jawad Ahmad, Baraq Ghaleb, Sana Ullah Jan, William J. Buchanan Abstract The rapid growth of cloud computing and data-driven applications has amplified privacy concerns, driven by the increasing demand to process sensitive data securely. Homomorphic encryption (HE) has become a vital solution for addressing these concerns by enabling computations on encrypted data without revealing its contents. This paper provides a comprehensive evaluation of two leading HE libraries, SEAL and OpenFHE, examining their performance, usability, and support for prominent HE schemes such as BGV and CKKS. Our analysis highlights computational efficiency, memory usage, and scalability across Linux and Windows platforms, emphasizing their applicability in real-world scenarios. Results reveal that Linux outperforms Windows in computation efficiency, with OpenFHE emerging as the optimal choice across diverse cryptographic settings. This paper provides valuable insights for researchers and practitioners to advance privacy-preserving applications using FHE. omomorphic encryption, HE libraries, cross-platform, resource utilization, privacy-preserving 1 Introduction Cloud computing has revolutionized modern computing by providing network access through cloud servers, which are remote storage systems that facilitate data processing and reduce the computational burden on local networks. While cloud servers offer significant advantages in terms of efficiency and scalability, their involvement also introduces critical security concerns, as sensitive data is entrusted to third-party entities. [ 1 ] . To address the security concerns associated with cloud computing, encryption techniques are employed to transform sensitive data into an unreadable form, ensuring its protection during transmission and storage. While essential for securing data during transmission and storage, traditional encryption techniques do not not fully address the privacy concerns in cloud computing as they typically require data to be decrypted before any computation can occur, exposing sensitive information to potential vulnerabilities during processing. This limitation underscores the need for more advanced encryption methods, such as Homomorphic Encryption (HE), which enables computations to be performed directly on encrypted data without the need for decryption. HE resolves the privacy challenges by allowing data to remain confidential throughout its lifecycle, from storage to computation. Figure 1 illustrates the distinction between conventional cloud computing and HE-based computing, emphasizing the enhanced security and privacy protection provided by HE. Figure 1: Architectural Difference between Cloud and HE-based Computing Homomorphic Encryption (HE) has made significant progress since its inception, evolving from a concept that allowed only simple arithmetic operations on encrypted data to a powerful tool for secure computations on large-scale data. Initially, HE was limited to performing either addition or multiplication on encrypted data, restricting its practical applicability. However, the field underwent a major breakthrough with Gentry’s introduction of Fully Homomorphic Encryption (FHE) in 2009, which enabled both addition and multiplication operations on encrypted data simultaneously. This advancement opened the door to more complex and scalable computations, making HE suitable for a broader range of real-world applications [ 2 ] . In his research, he addressed the limitation of initial HE by introducing both types of homomorphic operations on encrypted data. Despite its theoretical promise, FHE faced challenges in practical implementation due to its high computational complexity. In response, several specialized libraries, such as Microsoft SEAL and OpenFHE, were developed, each offering efficient implementations of different FHE schemes, including the BGV and CKKS schemes. These libraries have made it possible to perform precise integer computations (BGV) or approximate floating-point operations (CKKS), further expanding the versatility and feasibility of HE in privacy-preserving computing. As the adoption of Homomorphic Encryption (HE) grows, it becomes essential to thoroughly analyze and evaluate the available HE libraries. Indeed, each library, such as Microsoft SEAL and OpenFHE, offers different trade-offs in terms of computational efficiency, ease of use, and support for various HE schemes. Therefore, a comprehensive evaluation is necessary to identify the strengths and limitations of these libraries in real-world scenarios. This evaluation will help researchers, developers, and organizations make informed decisions about which library best meets their specific needs, whether it be for secure data processing in cloud environments or for implementing scalable privacy-preserving systems. Furthermore, analyzing these libraries will provide valuable insights into their optimization and potential areas for improvement, ensuring that HE can be effectively deployed in practical, large-scale applications. 1. How do FHE libraries perform across different configuration settings? 2. How do varying HE parameters impact the performance of FHE libraries? 3. How does the choice of platform (Linux vs. Windows) influence the performance of FHE libraries? 4. How do different FHE schemes (e.g., BGV vs. CKKS) perform across various FHE libraries and OS platforms? Hence, our study introduces a novel benchmark for evaluating the performance of Microsoft SEAL and OpenFHE, focusing on the BGV and CKKS schemes. We specifically examine two key FHE operations—homomorphic addition and multiplication. Additionally, we conduct a cross-platform analysis to assess the performance of these libraries in various environments. 2 Background Practical implementations of Homomorphic Encryption (HE) have faced several challenges, including mathematical complexity and increased resource demands. To address these limitations, various encryption schemes have been proposed, such as RSA, ElGamal, and Paillier. In 1977, Rivest-Shamir-Adleman (RSA) proposed a practical method of HE using multiplication functions on encrypted messages, m_{1} , and m_{2} , such as: \text{Enc}(m_{1})\cdot\text{Enc}(m_{2})=\text{Enc}(m_{1}\cdot m_{2}) (1) Similarly, Paillier introduced an additive version of RSA’s scheme where the operation is: \text{Enc}(m_{1})+\text{Enc}(m_{2})=\text{Enc}(m_{1}+m_{2}) (2) However, both RSA and Paillier schemes were limited to performing only one type of homomorphic operation—either addition or multiplication [ 3 ] , [ 4 ] . This single-operation behavior of the aforementioned schemes is known as partial homomorphic encryption (PHE) and has limited their applications in practical scenarios [ 5 ] . To overcome the limitations of PHE, Craig Gentry introduced the first fully homomorphic encryption (FHE) scheme [ 6 ] , which allowed both addition and multiplication operations on encrypted messages. For example, it enabled expressions like: \text{Enc}(m_{1}),\text{Enc}(m_{2})=\text{Enc}(m_{1}^{2}+m_{2}\cdot m_{1}) (3) Gentry’s FHE scheme is based on lattice-based cryptography and bootstrapping procedures. Lattice-based cryptography ensures data privacy by using complex mathematical problems such as Learning with Errors (LWE) whereas Bootstrapping is a procedure that refreshes ciphertexts (encrypted messages) to minimize the noise growth during encryption. Table 1: Focused Comparison of BGV and CKKS Aspect BGV CKKS Primary Use Case Integer values Floating-point values Arithmetic Type Modular arithmetics Approximate arithmetics Key Features Batching technique Scaling factor Applications Financial systems, cryptographic protocols. Machine learning, data analytics. 2.1 FHE parameters The adoption and effectiveness of Fully Homomorphic Encryption (FHE) in cryptographic applications depend significantly on the choice and optimization of several key parameters. These parameters directly influence the performance, security, and practicality of FHE schemes. Crucial parameters include the polymodulus degree, ciphertext modulus, plaintext modulus, scaling factor, and operation depth, all of which must be carefully chosen to balance security requirements and computational efficiency in FHE implementations [ 7 ] , [ 8 ] . Polymodulus Degree. The polymodulus degree, denoted by n , determines the size of the polynomial ring used in the encryption process. It is critical in defining the complexity and security of the encryption scheme. The choice of n impacts both the computational overhead and the level of security against attacks. In ring theory, a polynomial ring R is constructed using polynomials f(X) , where the coefficients belong to certain rings, and