Celestial sources for random number generation

Authors

Erin Chapman, Digital Forensics Research Laboratories, Institute for Radio Astronomy and Space Research, Auckland University of TechnologyFollow
Jerina Grewar, Digital Forensics Research Laboratories, Institute for Radio Astronomy and Space Research, Auckland University of TechnologyFollow
Tim Natusch, Digital Forensics Research Laboratories, Institute for Radio Astronomy and Space Research, Auckland University of Technology

Document Type

Conference Proceeding

Comments

Chapman, E., Grewar, J., & Natusch, T. (2016). Celestial sources for random number generation. In Johnstone, M. (Ed.). (2016). The Proceedings of 14th Australian Information Security Management Conference, 5-6 December, 2016, Edith Cowan University, Perth, Western Australia. (pp.5-13).

Abstract

In this paper, we present an alternative method of gathering seed data for random number generation (RNG) in cryptographic applications. Our proposed method utilises the inherent randomness of signal data from celestial sources in radio astronomy to provide seeds for RNG. The data sets were collected from two separate celestial sources, and run through the SHA-256 algorithm to deskew the data and produce random numbers with a uniform distribution. The resulting data sets pass all tests in the NIST Statistical Test Suite for random data, with a mean of 98.9% of the 512 total bitstreams from the two sources passing all tests in the NIST suite, as well as further testing in R. These results are on par with the control set generated using Java’s SecureRandom function. An explanation of the sources, the data processing and detailed results of each of the tests are presented.

DOI

10.4225/75/58a6975133e06