Mechanism for Detecting Domain Name System based Denial of Service Attacks

Abstract

The Domain Name System (DNS) is a critical component of the internet infrastructure, responsible for translating human-readable domain names into IP addresses. However, the DNS is vulnerable various attacks like DNS cache poisoning, DNS tunneling, denial of service (DoS), etc. Thus, an effective attack detection mechanism is required to prevent the malicious entry in the DNS. In this article, an Elman Neural Network-based attack detection mechanism was proposed to predict the normal and malicious traffic in DNS system. The proposed model utilizes Recursive Feature Elimination (RFE) approach to extract and select most relevant features to train the ENN model. The proposed work predicts the incoming network traffic as normal or malicious based on the trained feature set. Furthermore, an alert notification module was designed to notify the administrator about the entry of attack. The proposed model was trained, tested and validated with the ICS DNS dataset and the outcomes are estimated. The developed model earned greater performances of 99.89% accuracy, 99.76% precision, 99.59% recall, and 99.68% f-measure. Furthermore, the estimated outcomes are compared with some recent optimization and deep learning-based attack detection techniques. From the comparative assessment, it is observed that the performances are improved in the proposed technique compared to existing algorithms.