Advancing DDoS Attack Detection Using Machine Learning Strategies

Abstract

A vast amount of data is generated and saved on the cloud and other virtual storage systems as a result of the advancement of technologies like networking, cloud computing, and the Internet of Things (IoT) Massive volumes of data are sent and received over public or private networks, and mobile networks, Internet of Things networks, and other physical networks are essential for storing, sending, and analyzing that data. However, as a result of this rapid advancement in networking, cyberattacks have been observed on these networks, posing serious problems to the data security of these networks. Distributed Denial of Service (DDoS) attacks are the most common cyberattacks seen on any network, according to a recent survey. A network is inundated with DDoS attacks, which prevents the network from offering services to the authorized users. A Convolutional Neural Network-Visual Geometry Group (CNN-VGG) model developed in this study identifies DDoS attacks on any network. In proposed approach, CIC-IDS2017 (Intrusion Detection System) dataset is transformed into images, the model is trained and validated using these images. The model's performance is evaluated using measures like F1 Score, Accuracy, Precision, and Recall. The VGG model has proved a 92% accuracy rate in detecting DDoS attacks.