An algebraic attack on the improved summation generator with 2-bit memory

  • Authors:
  • Daewan Han;Moonsik Lee

  • Affiliations:
  • NSRI, 161 Gajeong-dong, Yuseong-gu, Daejeon 305-350, Republic of Korea;Department of Mathematics, Korea Air Force Academy, P.O. Box 335-2, Cheongwon, Chungbuk 363-849, Republic of Korea

  • Venue:
  • Information Processing Letters
  • Year:
  • 2005

Quantified Score

Hi-index 0.89

Visualization

Abstract

Recently algebraic attacks on stream ciphers have received much attention. In this paper we apply an algebraic attack to the improved summation generator with 2-bit memory, which was presented by Lee and Moon in order to give the original summation generator correlation immunity. We show that the initial state of the generator can be recovered within O(n^5^.^6) bit operations from O(n^2) regular output bits, where n is the total length of LFSRs. We could recover the initial key bits in practice within 3 minutes on a PC even for the case n=256. Our result is a good example that shows how powerful algebraic attacks are in the analysis of stream ciphers.