Multi-Class DDoS Attack Classification using Deep Neural Networks (DNN): A Critical Analysis

Abstract

The Multi-class Distributed Denial of Service (DDoS) attack classification using deep neural networks includes different types of network floods into specific types with high accuracy. A DDoS attack is when mischievous actors flood a server or network with fake traffic so real users cannot access the service. This emphasizes the use of intelligent systems for detection with the help of Machine Learning and Deep Learning. Such intelligent systems have overcome the problem of identifying huge volumes of malicious traffic among the legitimate network traffic and can handle high dimensional and complex data. This paper addresses need to develop, implement, and stringently evaluate a resource-efficient deep neural network model on the effective multi-class classification of 12 distinct types of DDoS attacks using the realistic CIC-DDoS2019 dataset in order to contribute to robust real-time threat detection in modern networks. The paper presents critical analysis which is beyond mere documentation to quantitative comparison and qualitative assessment, focusing on the trade-offs between performance, efficiency, and operational limitations. The proposed methodology has seven phases: data collection, data preparation, model design, model training, model evaluation, model validation, and model deployment. The core findings and achievements of the research are based on the methodology executed, have confirmed that the shift from the simple ML models to the more complex DL architectures, namely DNN, was well -grounded