A Queue Model to Detect DDos Attacks

Shuang Hao,Hua Song,Wenbao Jiang,Yiqi Dai

DOI: https://doi.org/10.1109/iscst.2005.1553301

2005-01-01

Abstract:With the development of network communication and collaboration, distributed denial-of-service (DDos) attack increasingly becomes one of the hardest and most annoying network security problems to address. In this paper, we present a new framework to detect the DDos attacks according to the packet flows of specific protocols. Our aim is to detect the attacks as early as possible and avoid the unnecessary false positive. A Gaussian parametrical mixture model is utilized to estimate the normal behavior and a queue model is adopted for detecting the attacks. Experiments verify that our proposed approach is effective and has reasonable accuracy.