Side Channel Analysis in Homomorphic Encryption

Baraq Ghaleb, William J Buchanan · 2025

Read Side Channel Analysis in Homomorphic Encryption by Baraq Ghaleb, 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-2505-11058.txt.

Introduction

Baraq Ghaleb and William Buchanan surveyed in 2025 how homomorphic encryption, which lets computers compute on data while it stays encrypted, can leak secrets through side channels. Side channels exploit implementation quirks like timing differences, a flaw class that has bitten older ciphers before. Fully homomorphic schemes rest on lattice mathematics, whose implementations carry their own leakage risks.

The paper sorts side-channel attacks on these systems into a clear categorization. Case studies show what concrete implementations reveal, from key material to plaintext hints. Countermeasures are reviewed for each weakness. The authors warn that mathematical strength alone does not secure a deployment. As encrypted computation moves into machine learning services, they argue, implementation hygiene deserves the same scrutiny as the cryptography itself.

Catalog record

CREATORBaraq Ghaleb, William J Buchanan
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-11058

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Side Channel Analysis in Homomorphic Encryption Baraq Ghaleb, William J Buchanan Abstract Homomorphic encryption provides many opportunities for privacy-aware processing, including with methods related to machine learning. Many of our existing cryptographic methods have been shown in the past to be susceptible to side channel attacks. With these, the implementation of the cryptographic methods can reveal information about the private keys used, the result, or even the original plaintext. An example of this includes the processing of the RSA exponent using the Montgomery method, and where 0’s and 1’s differ in their processing time for modular exponentiation. With FHE, we typically use lattice methods, and which can have particular problems in their implementation in relation to side channel leakage. This paper aims to outline a range of weaknesses within FHE implementations as related to side channel analysis. It outlines a categorization for side-channel analysis, some case studies, and mitigation strategies. \AppendGraphicsExtensions .gif 1 Introduction To enable secure processing of encrypted data, we require encryption schemes that support computation on ciphertexts. This can be achieved through Partial Homomorphic Encryption (PHE) or Fully Homomorphic Encryption (FHE). Encryption has been applied over-the-air and at-rest , but rarely during in-process computations. To achieve this, we require encryption schemes that support computation on ciphertext. This is where Homomorphic Encryption (HE) comes into play, enabling computations on encrypted data without decryption. HE can be categorized into Partial Homomorphic Encryption (PHE) and Fully Homomorphic Encryption (FHE). PHE supports a limited set of arithmetic operations, while FHE enables arbitrary computations on encrypted data. The PHE methods include RSA, ElGamal, Paillier [ 1 ] , Exponential ElGamal, Elliptic Curve ElGamal [ 2 ] , Paillier [ 1 ] , Damgard-Jurik [ 3 ] , Okamoto–Uchiyama [ 4 ] , Benaloh [ 5 ] , Naccache–Stern [ 6 ] , and Goldwasser–Micali [ 7 ] . Overall, we can use RSA and ElGamal for multiplicative homomorphic encryption; Paillier, Exponential ElGamal, Elliptic Curve ElGamal, Damgard-Jurik, Okamoto–Uchiyama, Benaloh and Naccache–Stern for additive homomorphic encryption; and Goldwasser–Micali for XOR homomorphic encryption. FHE, on the other hand, is typically based on lattice-based cryptography with modern implementations leverage libraries such as such as SEAL [ 8 ] and OpenFHE [ 9 ] . While FHE introduces innovative approaches to privacy-preserving computation, its practical implementation remains relatively new. Although the underlying cryptographic methods are theoretically secure, real-world deployments often face challenges, particularly in mitigating side-channel vulnerabilities. These vulnerabilities can unintentionally leak information about encryption keys or even the original plaintext data, posing significant security risks. This paper presents a comprehensive review of HE side-channel threats, highlighting key attack vectors and analyzing real-world case studies to assess their impact and potential countermeasures.. 1.1 Side-channel analysis Side-channel analysis (SCA) refers to a category of attacks that exploit information leaked from the physical implementation of cryptographic algorithms, rather than their mathematical weaknesses [ 10 ] [ 11 ] . Unlike traditional cryptanalysis, SCA leverages observable characteristics of a system during operation such as timing variations, power consumption, or electromagnetic emissions [ 12 ] [ 13 ] as depicted in Figure 1 . These leaks can provide insights into sensitive data enabling attackers to bypass algorithmic defenses. For example, during the encryption process in the RSA algorithm, the ciphertext is generated by: c\equiv m^{e}\pmod{n} (1) where c , m , e , and n represent ciphertext, plaintext, key, and the product of p and q , respectively. In practical implementations, modular exponentiation is commonly performed using the ”Square and Multiply” algorithm, as shown in Algorithm 1 . This method scans the exponent e bit by bit from left to right, performing an additional modular multiplication whenever a bit is set to 1. This variation causes detectable timing and power fluctuations, making it vulnerable to side-channel attacks where careful measurements and analysis can recover the secret key, bit by bit, as depicted in Figure 2 . Figure 1: Possible side channels of a cryptographic device during an encryption [ 13 ] . Algorithm 1 Exponentiation by Square-and-Multiply 1: x\leftarrow 1 2: for each bit of e from left to right do 3: x\leftarrow x^{2}\pmod{n} 4: if current bit of e is 1 then 5: x\leftarrow x\cdot m\pmod{n} 6: end if 7: end for 8: return x Figure 2: Observing RSA key bits using power analysis: The left peak shows the power consumption during the squaring-only step, the right (broader) peak shows the multiplication step, allowing exponent bits 0 and 1 to be distinguished [ 14 ] . Homomorphic encryption (HE) faces unique security challenges due to side-channel vulnerabilities which arise for several reasons including: • Computational Intensity: HE operations are highly computationally intensive, increasing exposure to side-channel analysis due to longer processing time [ 15 ] . • Complex Operations: HE schemes often involve complex mathematical operations with varying execution patterns that depending on the encrypted data or key material. These variations can result in distinguishable side-channel signatures [ 15 ] . • Cloud Computing Risk: HE is frequently deployed in cloud environments, introducing additional attack vectors as malicious co-located processes may exploit leakage through shared hardware resources [ 16 ] . 2 Background In 1978, Rivest, Adleman, and Dertouzos [ 17 ] were the first to introduce the concept of homomorphic operations, demonstrating its feasibility using the RSA cryptosystem. This early method supported multiply and divide operations [ 18 ] , but lacked the ability to perform addition and subtraction. Over time, HE encryption evolved into two main categories: Partially Homomorphic Encryption (PHE), which supports a limited set of arithmetic operations, and Fully Homomorphic Encryption (FHE), which enables unrestricted addition, subtraction, multiplication, and division. Since Gentry introduced the first FHE scheme in 2009 [ 19 ] , homomorphic encryption has evolved through four main generations: • 1st generation: Gentry’s method uses integers and lattices [ 20 ] including the DGHV method. • 2nd generation. Brakerski, Gentry and Vaikuntanathan’s (BGV) and Brakerski/ Fan-Vercauteren (BFV) leveraging the Ring Learning With Errors (RLWE) aproach [ 21 ] . These methods are similar to each other, and there is only a minor difference between them. They are generally used in applications with small integer values. • 3rd generation: These include DM (also known as FHEW) and CGGI (also known as TFHE) and support the integration of Boolean circuits for small integers. • 4th generation: CKKS (Cheon, Kim, Kim, Song) designed for efficient computations on floating-point numbers and can be applied to machine learning applications as it can implement logistic regression methods and other statistical computation [ 22 ] . 2.1 Public key or symmetric key Homomorphic encryption can be implemented either with a symmetric key or an asymmetric (public) key. With symmetric key encryption, we use the same key to encrypt as we do to decrypt, whereas, with an asymmetric method, we use a public key to encrypt and a private key to decrypt. In Figure 3 we use asymmetric encryption with a public key ( pk ) and a private key ( sk ). With this Bob, Alice and Peggy will encrypt their data using the public key to produce ciphertext, and then we can operate on the ciphertext using arithmetic operations. The result can then be revealed by decrypting with the associated

Read the complete text