Performance and Storage Analysis of CRYSTALS - Kyber (ML-KEM) as a Post - Quantum Replacement for RSA and ECC
Nicolas Rodriguez Alvarez, Fernando Rodriguez Merino · 2025
Read Performance and Storage Analysis of CRYSTALS - Kyber (ML-KEM) as a Post - Quantum Replacement for RSA and ECC by Nicolas Rodriguez Alvarez, Fernando Rodriguez Merino (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-2508-01694.txt.
Introduction
Nicolas Rodriguez Alvarez tested in 2025 whether the post-quantum cipher Kyber can stand in for RSA and elliptic-curve codes on plain hardware. Quantum machines inch toward the scale where Shor's method breaks today's codes. Kyber, standardized by NIST in 2022, is a lead replacement, yet slow uptake could leave systems exposed for years, as past hash upgrades did. His tests ran several Kyber setups using only built-in chip speedups.
Kyber held strong security against quantum attack while keeping speeds fair for most uses. No special hardware was needed beyond what makers already ship. The author pitches the cipher as ready for wide use and warns that delay, not the math, is the chief risk.
Catalog record
| CREATOR | Nicolas Rodriguez Alvarez, Fernando Rodriguez Merino |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-01694
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Performance and Storage Analysis of CRYSTALS - Kyber (ML-KEM) as a Post - Quantum Replacement for RSA and ECC Nicolas Rodriguez Alvarez, Fernando Rodriguez Merino Abstract The steady advancement in quantum computer error correction technology has pushed the current record to 48 stable logical qubits, bringinga us a little closer to machines capable of running Shor’s algorithm at scales that threaten RSA and ECC cryptography. While the timeline for developing such quantum computers remains uncertain, the cryptographic community must prepare for the transition to quantum-resistant algorithms. CRYSTALS-Kyber, standardized by NIST in 2022, represents a leading post-quantum cryptographic solution, but widespread adoption faces significant challenges. If this migration follows patterns similar to the SHA-1 to SHA-2 transition, organizations may experience prolonged periods of vulnerability, with substantial security and economic consequences. This study evaluates Kyber’s practical viability through performance testing across various implementation schemes, utilizing only standard built-in processor acceleration features (AES-NI, ASIMD) without any specialized hardware additions. Our findings demonstrate that Kyber provides robust security guarantees against quantum attacks while maintaining acceptable performance profiles for most contemporary applications, utilizing only commodity hardware with manufacturer-provided acceleration capabilities. 1 Introduction The theoretical foundations established by Shor and Grover [ Sho94 , Gro96 ] have evolved from academic concepts to practical concerns as quantum computing hardware continues to advance. While experts debate the timeline for achieving fault-tolerant quantum computers capable of running Shor’s algorithm at scale, the cryptographic community faces an urgent imperative: the transition to quantum-resistant algorithms cannot wait for quantum computers to become operational [ Glo24 , Gid25 ] . Historical precedents in cryptographic transitions offer sobering lessons about the challenges ahead. The migration from SHA-1 to SHA-2, initiated in 2005 following the discovery of collision vulnerabilities [ WYY05 , SBK+17 , Nat15 ] , took over a decade to complete [ PS ] , with many organizations maintaining vulnerable systems well beyond recommended timelines [ PS ] . This prolonged transition period exposed numerous systems to security risks and highlighted the substantial economic and operational costs associated with delayed cryptographic upgrades. Suppose the transition to post-quantum cryptography follows similar patterns. In that case, organizations may face extended periods of vulnerability to quantum attacks, with potentially catastrophic consequences for digital infrastructure, financial systems, and national security. In response to this challenge, the National Institute of Standards and Technology (NIST) initiated a comprehensive standardization process for post-quantum cryptographic algorithms, culminating in the selection of CRYSTALS-Kyber as the primary Key Encapsulation Mechanism (KEM) standard in 2022 [ Nat24 ] . Unlike RSA and ECC, which derive their security from number-theoretic problems vulnerable to quantum attacks, Kyber’s security relies on the learning with errors (LWE) problem over module lattices—a mathematical foundation believed to be resistant to both classical and quantum computational attacks [ BDK+18 ] . However, theoretical quantum resistance alone does not guarantee practical adoption. The success of any cryptographic standard depends critically on its performance characteristics, storage requirements, and compatibility with existing hardware infrastructure. Previous post-quantum proposals have faced significant barriers to adoption due to excessive computational overhead, prohibitive key sizes, or requirements for specialized hardware acceleration. To address these concerns and evaluate Kyber’s practical viability as a replacement for current cryptographic standards, this study conducts a comprehensive performance analysis across multiple architectures and implementation scenarios. This research contributes to the post-quantum cryptography transition by providing empirical evidence of Kyber’s performance characteristics under realistic deployment conditions. Our evaluation methodology utilizes standard hardware acceleration features available in commodity processors (AES-NI, AVX2, ASIMD) without requiring specialized, quantum-resistant hardware additions, ensuring that our findings accurately reflect the performance organizations can expect during real-world deployment. By comparing Kyber’s performance against equivalent-security implementations of RSA-7680 and SECP384R1 (ECC) across both x86_64 and ARM64 architectures, we establish benchmarks that inform migration planning and risk assessment for organizations preparing for the post-quantum era. 2 Background 2.1 Classical Cryptographic Schemes Modern public-key cryptography relies on two primary mathematical foundations that are vulnerable to quantum attacks. RSA (Rivest-Shamir-Adleman) cryptography derives its security from the computational intractability of the integer factorization problem—the difficulty of decomposing large composite numbers into their constituent prime factors. This problem becomes exponentially harder as key sizes increase, making RSA-2048 and RSA-4096 computationally infeasible to break with classical computers within reasonable timeframes [ LV01 ] [ RSA78 ] . Elliptic Curve Cryptography (ECC) offers an alternative approach based on the elliptic curve discrete logarithm problem (ECDLP) [ Mil07 ] . ECC achieves equivalent security levels to RSA with significantly smaller key sizes—a 256-bit ECC key provides comparable security to a 3072-bit RSA key. The SECP384R1 curve, standardized by the Standards for Efficient Cryptography (SEC), represents a widely deployed ECC implementation offering 192-bit security strength according to NIST guidelines [ Bar06 ] . Both RSA and ECC implementations typically employ hybrid encryption schemes that combine asymmetric and symmetric cryptography. In these systems, a Key Encapsulation Mechanism (KEM) securely exchanges a symmetric key using public-key methods. At the same time, a Data Encapsulation Mechanism (DEM) handles bulk data encryption using faster symmetric algorithms, such as AES or ChaCha20, which are theoretically unbreakable. 2.2 Quantum Algorithms Shor’s algorithm, formulated by Peter Shor in 1994, marked a milestone in quantum computing theory by showing that a sufficiently large, error-free quantum computer could factor large integers efficiently. Whereas the best-known classical algorithms run in subexponential time, Shor’s algorithm runs in “polynomial” time, approximately O(log(N)^{3}) depending on implementation details, making the cryptographic keys based on large - N factorization effectively breakable in negligible time (RSA and ECC) [ Sho97 ] . Meanwhile, Grover’s algorithm, introduced by Lov Grover in 1996, provides a quadratic speedup for unstructured search, reducing the classical cost of O(N) to O(\sqrt{N}) . Although not initially intended for factorization, Grover’s amplitude - amplification technique can be used to optimize specific subroutines within Shor’s method or to accelerate searches among partial solutions generated by Shor’s quantum circuit [ Gro96 ] . In theory, combining Shor’s and Grover’s algorithms could optimize the number of iterations and resource usage. 2.3 “Standard” hardware accelerations In this study, the default hardware accelerations provided by the CPU were utilized. The following section provides a comprehensive explanation of these accelerations. 2.3.1 Intel Intel® Advanced Encryption Standard Instructions ( AES-NI ): This hardware acceleration provides a speedup of 3 to 10x over an entirely software implementation using AES [ Int ] . Intel® Advanced Vector Extensions ( AVX/AVX2 ): Intel’s vector instruction set for SIMD vector operations [ Cor21 ] . The Kyber implementation used in this study leverages AVX2 instructions to accelerate its core lattice-based computations, resulting in significant performance improvements. Intel® Secure Key ( RDRAND/RDSEED ): On - chip, NIST - certified random number generator instructions for high - quality entropy source in key generation [ Cor21 ] Intel® Carry - Less Multiplication ( PCLMULQDQ ): Provides a single-cycle, hardware - accelerated carry - less 64x64 - bit multiply, used in GCM and other Galois - field operations [ Cor21 ] . 2.3.2 ARM AES : ARM v8 Cryptography Extensions add AESE/AESD/AESMC/AESIMC instructions for single - round encryption/decryption and key - schedule support [ ARM14 ] . SHA1/SHA2 : SHA1C / SHAP / SHAM / SHA256H / SHA256SU instructions accelerate SHA - 1 and SHA - 224/256 hashing