DDoS detection using CURE clustering algorithm with outlier removal clustering for handling outliers

DoS (Denial of Service) and DDoS (Distributed Denial of Service) is an anomalous traffic phenomena that is need serious attention. In the previous research has already been discussed on traffic anomaly detection based on clustering, with a hierarchical clustering algorithm method. In this paper, we introduce a method of network traffic anomaly (DDoS) detection using modernization of the traditional hierarchical clustering algorithm that is CURE clustering algorithm. CURE has advantages in the case of outliers. We modify the algorithm using outlier removal clustering (ORC) in terms of dealing with outliers. We apply the mechanism to detect and remove outliers from the specified clusters. We perform the outlier elimination scheme in two phase and do the removal at the point which detected as outlier. We also give an analysis and results of the proposed method.