Detecting Network Traffic-based Attacks Using ANNs

Abstract

Nowadays, data security is a significant challenge for computer networks, especially on internet-based systems and the internet of things (IoT). Many possible network attacks and intrusions need to stop and treat, but the first step is to stop the attack to discover it and understand its type. More specifically, active ones such as Denial of Service (DOS), Masquerade, Replays, Penetration, Placement, and unauthorized access. An attractive and practical field to satisfy attack detection and prediction is Machine Learning (ML), which has techniques such as Artificial Neural Networks (ANNs) that take the data transmission request vectors and rely on them to classify the attacks. ANNs have many structure options so selected the most appropriate structure for the article context: the Feed-Forward Back-Propagation structure. Hence, introducing the ANN technique and applying it to an international dataset will discover how the experimental results would prove a significant acceptable accuracy of attack detection. Moreover, the article margin discussed two of the standard techniques for fighting the attacks to give recommendations for best practices, which are the Digital Signature and the Cryptography functions, these methods that can decrease and harden the attacks, then the role of the ML techniques would be more specific and determined..