Securing Distributed IoT Routing Networks Against DDoS Attacks Using Intelligent Machine Learning Techniques

Abstract

The rampant deployment of the Internet of Things (IoT) has increased the data traffic in the interconnected devices, which has also raised the cybersecurity concerns in IoT networks, especially the DDoS attacks targeting IoT. Conventional security approaches like password encryption/authentication to be broken are unsuitable for twarthmanaging these sophisticated, changing network threats, particularly in a distributed computing-based routing structure. This paper presents a holistic ML-driven framework that detects and mitigates DDoS attacks aimed at distributed IoT routing systems. The solution uses SVM, RF and DT supervised machine learning algorithms to classify malicious network features and increase the ability to detect intrusion mechanisms in real time. The models are based on historical network traffic data which are used to identify anomalous patterns and forecast future attack vectors. Performance evaluation is performed using important classification matrices such as confusion matrix, F1-score and AUC-ROC in order to ensure effective treatment for imbalanced datasets. Experimental results: The results were said to have been used on the Random Forest algorithm that gives ac-curacy of 99.2%, with 0.8% false positive rate, and 0.997 concordance index which is equivalent to the AUC-ROC score. The results substantiate the efficiency of self-learning intelligent machine learning based approach in hardening IoT routing networks for counteracting complex form of DDoS threats in the distributed domain.