A Flow Based Horizontal Scan Detection Using Genetic Algorithm Approach
Morteza Barati, Zahra Hakimi, Amir‐Homayoun Javadi · UCL Discovery (University College London) · 2013
An attacker has to 'scan' susceptible points of a network before attacking. There are several methods of detection of such behavior which are mostly based on thresholding. As the performance of these methods is highly dependent on the value of threshold, it is crucial to adjust this value appropriately. This adjustment is not always trivial. In this study we proposed a new method to optimize the parameters of the system using genetic algorithms (GA) based on network flows. Subsequently we compared our method with Snort. The results showed a superior performance as measured by the sensitivity index of d'.