Detection of unknown dos attacks by kolmogorov-complexity fluctuation

  • Authors:
  • Takayuki Furuya;Takahiro Matsuzaki;Kanta Matsuura

  • Affiliations:
  • Institute of Industrial Science, The University of Tokyo, Tokyo, Japan;Institute of Industrial Science, The University of Tokyo, Tokyo, Japan;Institute of Industrial Science, The University of Tokyo, Tokyo, Japan

  • Venue:
  • CISC'05 Proceedings of the First SKLOIS conference on Information Security and Cryptology
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Detection of unknown Denial-of-Service (DoS) attacks is a hard issue. What attackers do is simply to consume a large amount of target resources. This simple feature allows attackers to create a wide variety of attack flows, and hence we must find a sophisticated general metric for detection. A possible metric is Kolmogorov Complexity (KC), a measure of the size of the smallest program capable of representing the given piece of data flows because DoS attacks, known or unknown, are anyway launched by computer programs. However, there are no established DoS-detection methods which make use of this possibility. And to make matters worse, it is well known that KC cannot be rigorously computed. In this paper, we compare three different KC estimation methods including a new proposal of our own, and propose a new DoS-detection method by monitoring fluctuation of KC differentials.