A Detection Techniques for Botnet Attacks in Internet of Things Networks: A Comprehensive Literature Review
The Botnet attacks are among the most common and dangerous cybersecurity threats in the Internet of Things (IoT) environments. This is because of the massive number of diverse, low-resource devices and weak security settings. Traditional botnet detection methods, such as signature-based and rule-based Intrusion Detection System (IDS), are still widely adopted, but they struggle to detect new botnet variants and unknown attacks. In recent years, Artificial Intelligence (AI) has improved the detection performance through identifying patterns in traffic behavior and features of network flows. This paper presents a comprehensive review of the IoT botnet detection approaches, which are classified as either a non-AI detection system (NAIDS-Botnet) or an AI detection system (AIDS-Botnet). It surveys recent studies and datasets and provides a comparative analysis of different approaches. The analysis shows that non-AI-based techniques typically achieve accuracies of 85% to 93%, while AI-based approaches generally report accuracies between 92% and 99%. It also discusses current challenges like data imbalance, lack of evaluation on real networks, concept drift, deployment issues, etc. Finally, the paper outlines the advantages of AI-based IDS and proposes future research directions.
Authors
- Krishan Kumar
- Sukhvir Singh
- Sudhanshu Angiras
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-13
- DOI
- https://doi.org/10.5281/zenodo.24200322
- Primary Topic
- Network Security and Intrusion Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00