Highly Responsive Distributed Denial-of-Service Attacks Detection by Using Real-Time Burst Detection Method

Abstract

The accuracy of statistical methods for detecting a distributed denial-of-service (DDoS) attack improves with increasing packet sequence number (window size). However, such methods tend to suffer from low responsiveness. An alternative approach, real-time burst detection, offers two advantages over traditional statistical methods. First, it can be used for real-time detection when a DDoS event is occurring, and second, it can be judged with less processing as information about events can be compressed, even if a large number of events occur. Here, we propose a highly response burst detection method for DDoS attacks, perform experiments to evaluate its effectiveness, and discuss its detection accuracy and processing performance.