Distributed mechanism in detecting and defending against the low-rate TCP attack

  • Authors:
  • Haibin Sun;John C. S. Lui;David K. Y. Yau

  • Affiliations:
  • Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shatin, NT, Hong Kong;Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shatin, NT, Hong Kong;Department of Computer Science, Purdue University

  • Venue:
  • Computer Networks: The International Journal of Computer and Telecommunications Networking
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we consider a distributed mechanism to detect and to defend against the low-rate TCP attack. The low-rate TCP attack is a recently discovered attack. In essence, it is a periodic short burst that exploits the homogeneity of the minimum retransmission timeout (RTO) of TCP flows and forces all affected TCP flows to backoff and enter the retransmission timeout state. When these affected TCP flows timeout and retransmit their packets, the low-rate attack will again send a short burst to force these affected TCP flows to enter RTO again. Therefore these affected TCP flows may be entitled to zero or very low transmission bandwidth. This sort of attack is difficult to identify due to a large family of attack patterns. We propose a distributed detection mechanism to identify the low-rate attack. In particular, we use the dynamic time warping approach to robustly and accurately identify the existence of the low-rate attack. Once the attack is detected, we use a fair resource allocation mechanism to schedule all packets so that (1) the number of affected TCP flows is minimized and, (2) provide sufficient resource protection to those affected TCP flows. Experiments are carried out to quantify the robustness and accuracy of the proposed distributed detection mechanism. In particular, one can achieve a very low false positive/negative when compare to legitimate Internet traffic. Our experiments also illustrate the the efficiency of the defense mechanism across different attack patterns and network topologies.