Security of cryptographic algorithms and protocols in most security systems is guaranteed by assuming the existence of ideal random number generators (RNGs). In cryptographically secure RNG, its randomness is derived by some noise sources. An entropy source consisted of multiple noise sources outputs non-IID (independent and identically distributed) bit-sequences in general. Markov chain is typical non-IID stochastic model and the minentropy has important applications for randomness extractors. In this study, we classify first-order Markov chain into three types and obtain the theoretical min-entropy rates of these, respectively. Moreover, we compare our theoretical results with some experimental results which are obtained from the min-entropy estimation procedure of non-IID track in NIST SP 800-90B.

Published in: International Conference on Information Society (i-Society 2018)

  • Date of Conference: 15-18 July 2018
  • DOI: 10.2053/iSociety.2018.0004
  • ISBN: 978-1-908320-92-6
  • Conference Location: Dublin, Ireland