An Intrusion-Tolerant Intrusion Detection Method Based on Real-Time Sequence Analysis

One of the most advanced research issues in network security is intrusion-tolerant intrusion detection, which has become another essential technique to protect computer systems and prevent the intrusion from generating a system failure. This paper presents a novel intrusion-tolerant intrusion detection method based on real-time sequence forecast analysis for network stream. We devise linear regression techniques to forecast network stream sequences. According to these, it´s helpful for us to analysis intruders´ behaviors and to recognize undesirable intrusions. We also provide recovery strategies to tolerate intrusion. Experiments on the http server demonstrate that our method outperform the others