The Optimistic Schemes of Cluster Analysis and k-NN Classifier Method in Detecting and Counteracting Learned DDoS Attack

Abstract

The creation of Internet has been materialized to help people become aware of different information and unleash them from the state of ignorance. However, its vast expansions turned out to be a threat at their individual premises wherein integrity, accessibility and confidentiality are oftentimes compromised. This paper concerns the optimistic schemes of detecting and counteracting learned DDoS attacks. We described approaches of cluster analysis and k-NN classifier method as effective tools to battle tremendous security threats i.e., malicious usage, attacks and sabotage. These schemes were tested using a set of benchmark data from KDD (Knowledge Discovery and Data Mining) designed by DARPA. Results are clear evidence that combinations of such schemes lead to have an efficient and accurate performance in detecting DDoS attacks.