Machine Learning-Based IoT-Botnet Attack Detection with Sequential Architecture.

Area of Science:

• Cybersecurity
• Machine Learning
• Internet of Things (IoT)

Background:

• The proliferation of Internet of Things (IoT) devices has led to a surge in cyber-attacks, with botnets being a primary threat.
• Existing IoT devices often lack the resources for robust security, and rule-based detection systems are vulnerable to circumvention.
• There is a critical need for efficient and effective botnet attack detection mechanisms tailored for resource-constrained IoT environments.

Purpose of the Study:

• To propose a novel machine learning (ML)-based framework for detecting botnet attacks in IoT environments.
• To develop a lightweight and high-performance detection system through an efficient feature selection approach.
• To validate the effectiveness of a sequential detection architecture for identifying botnet-based threats.

Main Methods:

• Development of a machine learning-based framework utilizing a sequential detection architecture.
• Implementation of an efficient feature selection method to create a lightweight detection system.
• Evaluation of the framework using three distinct ML algorithms: Artificial Neural Network (ANN), J48 decision tree, and Naïve Bayes.

Main Results:

• The proposed framework achieved an overall detection performance of approximately 99% for botnet attacks.
• The sequential architecture demonstrated effectiveness in identifying botnet-based attacks across different ML algorithms.
• The feature selection approach contributed to a lightweight yet high-performance detection system.

Conclusions:

• The proposed ML-based sequential detection framework is highly effective for identifying botnet attacks on IoT devices.
• The system's lightweight design and high accuracy make it suitable for resource-constrained IoT environments.
• The architecture is extensible, allowing for the integration of sub-engines to detect emerging or new types of cyber-attacks.