CLSTMNet Architecture: A CNN–LSTM-Based Hybrid Deep Learning Model for DDoS Attack Detection and Mitigation in Network Security

Abstract

Distributed denial-of-service (DDoS) attacks represent one of the most damaging cybersecurity threats to modern network systems. The impact of this attack causes server failure and creates complaints about service inconvenience from users, thus reducing the company’s reputation and trust; more crucially, it is the loss of revenue. Although intrusion detection systems (IDSs) and other conventional security mechanisms have been widely deployed, many advanced DDoS attacks continue to bypass these defenses due to their evolving and complex patterns. This study aims to provide a state-of-the-art strategy to identify denial-of-service (DDoS) attacks more precisely using machine learning (ML) calculations. Creation of a modern deep learning (DL) strategy identifies DDoS attacks more precisely by combining the two best DL calculations and comparing their execution by actualizing them on the most challenging dataset. This research applies a combination strategy of two DL calculations models, convolutional neural network (CNN) and long short-term memory (LSTM). These calculations are actualized on the Network Security Laboratory–Knowledge Discovery and Data Mining (NSL-KDD) dataset, which is considered the most challenging dataset for DDoS attack discovery. The results show that the modern DL strategy created in this consideration outperforms other state-of-the-art strategies in terms of precision and discovery rate. The combination of CNN and LSTM results in superior execution than either calculation alone. This implies that the modern DL strategy created in this consideration is a feasible approach to identify DDoS attacks with high precision.