Connections among nonlinearity, avalanche and correlation immunity

  • Authors:
  • Yuliang Zheng;Xian-Mo Zhang

  • Affiliations:
  • Department of Software and Information Systems, The University of North Carolina at Charlotte, 9201 University City Blvd, Charlotte, NC;Department of Computing, Macquarie University, Sydney, NSW 2109 North Ryde, Australia

  • Venue:
  • Theoretical Computer Science - Algorithms,automata, complexity and games
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Nonlinear Boolean functions play an important role in the design of block ciphers, stream ciphers and one-way hash functions. Over the years researchers have identified a number of indicators that forecast nonlinear properties of these functions. Studying the relationships among these indicators has been an area that has received extensive research. The focus of this paper is on the interplay of three notable nonlinear indicators, namely nonlinearity, avalanche and correlation immunity. We establish, for the first time, an explicit and simple lower bound on the nonlinearity Nf of a Boolean function f of n variables satisfying the avalanche criterion of degree p, namely, Nf ≥ 2n-1 -2n-1-(1/2)p. We also identify all the functions whose nonlinearity attains the lower bound. As a further contribution of this paper, we prove that except for very few cases, the sum of the degree of avalanche and the order of correlation immunity of a Boolean function of n variables is at most n-2. The new results obtained in this work further highlight the significance of the fact that while avalanche property is in harmony with nonlinearity, both go against correlation immunity.