CHAOS IN CRYPTOGRAPHY: THE ESCAPE FROM THE STRANGE ATTRACTOR
John M. Carroll, Jeff Solo, Perry T. Wong · Cryptologia · 1992
Modern cryptography has a voracious appetite for reproducible sequences of large numbers that possess random characteristics. Because a cipher system is vulnerable to attack if cryptanalysts can successfully guess the nature of its underlying pseudorandom number generators, it is helpful for cryptographers to have a large repertoire of them. Chaos theory provides many formulations that can be used as random generators. This paper describes one of these, the Lorenz attractor. The chaotic nature of the Lorenz system of equations makes it a good candidate for pseudo-random number generation. However, to ensure that its sequences are not serially autocorrelated, they must be modified so that the particle can escape from the field of the strange attractor. The modified system produces extremely long sequences with good random properties. The generating algorithm is fast and efficient. Significantly, several components are needed to specify the starting point, which increases resistance to cryptanalysis and admits a number of key management and authentication protocols.