Random Number Generation from Pulsars

Hayder Tirmazi · 2025

Read Random Number Generation from Pulsars by Hayder Tirmazi (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-2502-18430.txt.

Introduction

Tirmazi draws random numbers from the timing jitter of pulsars in twenty twenty-five. Pulsars flash with near clockwork gaps, from milli seconds to seconds. Tiny wobbles in those gaps carry true noise from space. Those wobbles can seed a random number generator.

Extractor steps then polish the raw bits toward full entropy. The source is modeled in a crypto frame and shown close to ideal uniform chance. Standard test suites, including the NIST battery, pass the output. The work casts dead stars as a public source of randomness. Their signals are free for anyone to record, yet hard to fake.

Catalog record

CREATORHayder Tirmazi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE19 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-18430

Browse the full library · About the Tower of Babel

Full text

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

Random Number Generation from Pulsars Hayder Tirmazi Abstract Pulsars exhibit signals with precise inter-arrival times that are on the order of milliseconds to seconds, depending on the individual pulsar. There are subtle variations in the timing of pulsar signals. We show that these variations can serve as a natural entropy source for the creation of Random Number Generators (RNGs). We also explore the effects of using randomness extractors to increase the entropy of random bits extracted from Pulsar timing data. To evaluate the quality of the Pulsar RNG, we model its entropy as a k k -source and use well-known cryptographic results to show its closeness to a theoretically ideal uniformly random source. To remain consistent with prior work, we also show that the Pulsar RNG passes well-known statistical tests such as the NIST test suite. keywords cryptographic randomness, random number generation, trngs, astronomy 1 Introduction Random number generators (RNGs) are a fundamental part of modern cryptography ( Katz and Lindell 2014 ) . They can be used to implement provably secure secret-key encryption schemes ( Katz and Lindell 2014 ; Pass and Shelat 2010 ) , digital signature schemes ( Katz and Lindell 2014 ) , and the key generation step of public-key encryption schemes such as RSA ( Rivest et al. 1978 ) and ElGamal 1985 . True Random Number Generators (TRNGS) use noise in physical processes as a source of randomness. As an example, Intel’s TRNG uses Johnson noise in resistors ( Jun and Kocher 1999 ) . Pseudo-Random Number Generators (PRNGs) are initialized with a seed and use algorithms to produce numbers that seem random to adversaries that do not know the seed and are restricted to performing all their computations in probabilistic polynomial time Pass and Shelat 2010 . The initial seed of a PRNG may be derived from a TRNG. Prior work on extracting randomness from astrophysical sources includes, in chronological order, hot pixels in astronomical imaging ( Pimbblet and Bulmer 2005 ) , radio astronomy signal data noise ( Chapman et al. 2016 ) , cosmic microwave background radiation spectra ( Lee and Cleaver 2017 ) , cosmic photon arrival times ( Wu et al. 2017 ) , and intrinsic flux density distribution of single pulsars ( Dawson et al. 2022 ) . (a) J0030+0451 NANOGrav (b) J0030+0451 EPTA (c) J1918-0642 NANOGrav (d) J1918-0642 EPTA Figure 1 : Timing Variations in \mu s for the J0030+0451 and J1918-0642 pulsars with Modified Julian Date (MJD) and Year plotted on the x axis Pulsar timing variations provide an alternative entropy source that is structured yet unpredictable. To the best of our knowledge, this paper is the first to investigate the variation in inter-arrival times of pulsar signals as a novel entropy source for cryptographic random number generation. Prior work on random number generation from astrophysical sources notably including Dawson et al. 2022 and Pimbblet and Bulmer 2005 has relied primarily on black-box statistical testing to evaluate randomness quality. There are well-known concerns ( Saarinen 2022 ) with relying solely on such statistical tests without a proper theoretical analysis of the entropy source. In fact, the statistical tests endorsed by NIST can be passed even by weak PRNGs ( Saarinen 2022 ) . Unlike prior work, our work also includes a theoretical analysis using known cryptographic techniques to complement our empirical findings. The rest of this paper is structured as follows. In Section 2 , we create a Pulsar RNG from observational data from two sources: the North American Nanohertz Observatory for Gravitational Waves (NANOGrav) ( Matthews et al. 2016 ) and the European Pulsar Timing Array (EPTA) ( EPTA et al. 2024 ) . Section 3 evaluates the Pulsar RNG using a cryptographic analysis and statistical tests. Section 4 provides discussion relevant to the viability of the Pulsar RNG. 2 Generating Random Bits We use measurement data from two pulsars, PSR J0030+0451 and PSR J1918-0642. These two pulsars are present in both the NANOGrav 9-year dataset release ( Jun and Kocher 1999 ) and the EPTA DR2 dataset release ( EPTA et al. 2024 ) . Our Pulsar RNG extracts timing residuals from these datasets using PINT ( Luo et al. 2021 ) v1.1.1. Let L be the list of pulsar residuals where L_{i} is the ith element. We first normalize the residual values to create list N in the usual way ( N_{i}=\frac{L_{i}-\min(L)}{\max(L)-\min(L)} ). We then investigate three quantification techniques on the list N to convert it to a list of random bits R . 1. A simple threshold: R_{i}=1 if N_{i}\geq 0.5 , otherwise R_{i}=0 2. 8-bit Gray coding ( Doran 2007 ) 3. Using the 8-bit Gray coded value as a seed for a SHA-512 hash ( Penard and Van Werkhoven 2008 ) Figure 2 shows the measured dataset entropy in bits per byte of these three quantification methods. Note that by entropy throughout this paper we mean information entropy, also known as Shannon entropy. We measure all dataset entropy results in this paper using the ent tool ( Walker 2008 ) . Figure 2 : Entropy of Different Quantification Methods for 2 Pulsars across EPTA and NANOGrav data Figure 3 : Normalized PSR J1918-0642 residuals on EPTA data (above) and NANOGrav data (below). Using threshold as a quantification method requires a careful choice of where to put the threshold based on each distribution. Our method of uniformly using \tau=0.5 as the threshold provides vastly different results for, as an example, the PSR J1918-0642 data on the EPTA dataset as opposed to the NANOGrav dataset. This is due to \tau=0.5 not providing an equal direction of the EPTA data. This can be verified in Figure 3 which shows normalized residuals ( N ) for both datasets. Notice that while the points on the NANOGrav data are roughly equally divided by a cutoff line at 0.5 (marked by a dotted line in the figure), most points in the EPTA dataset are below 0.5 . We next investigate the effect of three different randomness extractors on our results. Randomness Extractors are functions that take as input 1) a comparatively small uniformly-random seed and 2) a comparatively weak entropy source, for example, radioactive decay ( Walker 2001 ) or in our case Pulsar timing variation. Randomness Extractors output random bits that appear to computationally bound adversaries as being independent from the input entropy source and uniformly randomly distributed. Note that prior astrophysics-based RNG papers including Pimbblet and Bulmer 2005 refer to randomness extractors as debiasing or deskewing algorithms. We test two simple ad hoc randomness extractors, XOR-ing several subsequent bits ( Stipčević and Koç 2014 ) and Von Neumann 1963 . We also test a Randomness extractor based on SHAKE-256 from the SHA-3 family of cryptographic hash functions. Figure 4 shows our results. While using a cryptographic hash yields the highest entropy, it is interesting to note that even an ad-hoc random extractor like Von Neumann provides considerable entropy gains. Figure 4 : Entropy for different randomness extractors on data from PSR J0030+0451 (EPTA). 3 Evaluation A strict mathematical proof of absolute randomness is considered impossible ( Stipčević and Koç 2014 ) . To analyze TRNGs, we must rely on assumptions based on the fundamental

Read the complete text