[Paper Review] Divisibility, Smoothness and Cryptographic Applications

David Naccache, Igor E. Shparlinski|ArXiv.org|Oct 12, 2008

Analytic Number Theory Research73 references3 citations

TL;DR

This paper explores the role of smooth numbers—integers free of large prime factors—in cryptographic systems, demonstrating how number-theoretic properties of divisibility and smoothness underpin the security and efficiency of constructions like RSA, hash functions (e.g., VSH), and factoring algorithms. It establishes rigorous bounds on smooth number distributions and applies them to prove collision resistance and security guarantees in cryptographic primitives.

ABSTRACT

This paper deals with products of moderate-size primes, familiarly known as smooth numbers. Smooth numbers play a crucial role in information theory, signal processing and cryptography. We present various properties of smooth numbers relating to their enumeration, distribution and occurrence in various integer sequences. We then turn our attention to cryptographic applications in which smooth numbers play a pivotal role.

Motivation & Objective

• To analyze the cryptographic significance of smooth numbers and their distribution in integer sequences.
• To provide rigorous number-theoretic foundations for cryptographic constructions relying on divisibility and smoothness.
• To bridge gaps between analytic number theory and practical cryptography by linking smoothness to security proofs and algorithmic efficiency.
• To demonstrate how refined estimates on divisor functions and smooth number counts underpin modern cryptographic protocols.
• To show that small improvements in analytic number theory often yield major cryptographic advances, such as tighter security bounds and efficient algorithms.

Proposed method

• Uses Vinogradov’s big-O notation and asymptotic analysis to bound the count of smooth numbers up to a given limit.
• Applies results from Ford (2008) on the distribution of values of Euler’s totient function and the number of divisors in arithmetic progressions.
• Employs estimates on the stopping time of the Dixon factoring algorithm via Croot, Granville, Pemantle & Tetali (2006) to analyze algorithmic efficiency.
• Analyzes the multiplicative subgroup generated by the first $k$ primes modulo $N = pq$ to assess the security of the Very Smooth Hash (VSH) function.
• Leverages bounds on $y$-smooth numbers (e.g., $ ilde{ ho}(x,y)$) to prove that for almost all RSA moduli, the probability of a random message hashing to a given value is negligible.
• Combines classical results on the Chinese Remainder Theorem and Fast Fourier Transform (FFT) recursion on smooth $n$ to illustrate how smoothness enables sub-quadratic algorithms, though these are excluded from the main cryptographic focus.

Experimental results

Research questions

• RQ1 How common are integers whose prime factors are all below a given bound $b$, and how does this affect cryptographic security?
• RQ2 What is the distribution of values taken by Euler’s totient function $ au(n)$, and how does this relate to the security of RSA?
• RQ3 How can the smoothness of $p-1$ and $p+1$ for primes $p$ be used to define and analyze 'strong' primes?
• RQ4 What is the probability that a random message maps to a given hash value under the Very Smooth Hash (VSH) function, and how can this be bounded using smooth number theory?
• RQ5 To what extent do refined analytic number theory results (e.g., on divisor counts or smooth number counts) improve the security and efficiency of cryptographic algorithms?

Key findings

• The number of $y$-smooth integers up to $x$, denoted $ ilde{ ho}(x,y)$, is bounded using results from analytic number theory, with $ ilde{ ho}(x,y) o 0$ as $x/y o au$ for $ au > 1$, implying that most integers are not $y$-smooth.
• For almost all RSA moduli $N = pq$, the probability that a random $ ho$-bit message maps to a given hash value under VSH is negligible when $ ho$ is sufficiently large, due to the distribution of smooth numbers in the multiplicative group modulo $N$.
• The set of primes $p eq 2$ for which $p-1$ is $y$-smooth has density $O(x/y)$, implying that such primes are rare when $y$ is small relative to $x$.
• The bound on the number of divisors $d eq 0 mod k$ with $d eq a mod k$ is derived from Coppersmith’s method, which underpins attacks on RSA with partially known factors.
• The stopping time of the Dixon factoring algorithm is tightly bounded using results from Croot et al., showing that the algorithm runs in sub-exponential time when $n$ is smooth.
• The probability that $ au(p-1)$ is $y$-smooth is negligible for large $x$ and $y$, implying that most primes are $y$-strong, which supports the security of strong prime generation in RSA.