Anticipating Cyber Threats: Deep Learning Approaches for DDoS Attacks Forecasting

This paper investigates the application of deep learning (DL) techniques for DDoS attacks forecasting (i.e., predicting future attacks beforehand). We adopted a uniform modeling approach that allowed us to compare the performances of various DL algorithms, namely Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU). The results show that both short-term (few seconds ahead) and mid-term (up to 20 seconds ahead) DDoS traffic prediction with a very high accuracy were possible. In addition, the results showed that LSTM model outperforms both RNN and GRU, as well as machine learning algorithms such as Support Vector Machines (SVM), Random Forests (RF), and k-Nearest Neighbors (KNN).