Deep learning-driven architecture for effective DDoS attack detection

A Distributed Denial of Service (DDoS) attack is intended to hinder network assets, making them unavailable to approved users. Attacks have become increasingly common, posing significant threats to internet users. DDoS attacks seek, in most cases, to temporarily or even irreparably cripple a target's web presence. In contrast, cloud infrastructure continues to develop, with container technology allowing for efficient use of resources and expandable service delivery. The current work proposes a new scheme for DDoS detection using deep learning in conjunction with a math model for countermeasures. In its proposed algorithm, artificial intelligence methodologies such as Deep Convolutional Neural Networks (DCNNs) and an autoencoder have been added to enhance accuracy in detection. With use of existing databases, performance in relation to complex detection tools is similar, yet processing time is significantly reduced. It is argued that such an approach is specifically suitable for operational environments with restricted capabilities.