Predicting DDoS Attacks: How Deep Learning Could Give Defenders An Early Warning
Help Net Security, Thursday, September 25th, 2025
Distributed denial-of-service (DDoS) attacks remain one of the most common and disruptive forms of cybercrime. Defenders have traditionally focused on detecting these attacks once they are underway.
New research suggests that predicting DDoS attacks in advance may be possible, giving security teams a head start in planning their defenses.
A new study outlines an approach to forecasting DDoS activity using deep learning. The researchers from Universiti Malaya and Universiti Teknikal Malaysia Melaka analyzed 192,525 DDoS attacks that took place between 2019 and 2021. Their goal was to determine whether patterns in past activity could be used to forecast upcoming surges.