Evaluation on multivariate correlation analysis based denial-of-service attack detection system

In this paper, a Denial-of-Service (DoS) attack detection system is explored, where a multivariate correlation analysis technique based on Euclidean distance is applied for network traffic characterization and the principal of anomaly-based detection is employed in attack recognition. The effectiveness of the detection system is evaluated on the KDD Cup 99 dataset and the influence of data normalization on the performance of attack detection is analyzed in this paper as well. The evaluation results and comparisons prove that the detection system is effective in distinguishing DoS attack network traffic from legitimate network traffic and outperforms two state-of-the-art systems.