Multiprime Blum-Blum-Shub Pseudorandom Number Generator

Abstract

Blum-Blum-Shub (BBS) is a (probabilistically) secure pseudorandom bit/number generator which outputs a sequence by repeatedly reducing squares modulo the product of two Blum-primes. Our goal for this thesis is to modify the algorithm by using a modulus which is the product of three Blum-primes. We evaluate the effect of this modification using the suite of tests from National Institute of Standards and Technology (NIST). Previous research has evaluated the limit on the number of least important bits that can be extracted per iteration of the BBS algorithm while still maintaining the pseudorandom properties. In this paper, we go beyond the proposedlimit and compare the modified BBS with the original BBS using the NIST tests. This paper also discusses the cryptosystem based on the modified BBS as well as the original BBS. We use three metrics for the comparison of performance: the type of tests, the overall performance of sequences against NIST tests, and the time to generate sequences. Our test data shows that both versions performed ina similar manner when subjected to NIST tests. Furthermore, bit generation is significantly faster for sequences generated by taking the last 50 bits or more, while still maintaining pseudorandom properties.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2016
Accession Number
AD1030047

Entities

People

  • Bijesh Shrestha

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Biomedical
  • C4I

DTIC Thesaurus Topics

  • Algorithms
  • Computer Programming
  • Cryptography
  • Demographic Cohorts
  • Generators
  • Iterations
  • Mathematics
  • Number Theory
  • Numbers
  • Operating Systems
  • Prime Numbers
  • Probability
  • Random Walk
  • Sequences
  • Square Roots
  • Standards
  • Template Patterns

Fields of Study

  • Computer science
  • Mathematics

Readers

  • Computer Programming and Software Development.
  • Regression Analysis.