Transport and Application Layer DDoS Attacks Detection to IoT Devices by Using Machine Learning and Deep Learning...

Area of Science:

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

Background:

• The proliferation of Internet of Things (IoT) devices introduces significant security vulnerabilities due to a lack of standardized security protocols across diverse manufacturers.
• Denial of Service (DoS) and Distributed Denial of Service (DDoS) attacks are prevalent and critical threats to IoT networks, with a notable increase observed in advanced targeted attacks.

Purpose of the Study:

• To develop a novel Intrusion Detection System (IDS) for IoT environments to combat DoS and DDoS attacks.
• To address the class imbalance problem within the Bot-IoT dataset for improved attack detection.
• To evaluate the impact of feature sets and timestamps on prediction accuracy and performance.

Main Methods:

• Utilized the Bot-IoT dataset, employing Machine Learning and Deep Learning models for intrusion detection.
• Implemented three distinct feature sets for binary and multiclass classifications to mitigate feature dependencies.
• Conducted comprehensive experimentation, including time performance evaluation, to assess model effectiveness.

Main Results:

• Achieved an average accuracy exceeding 99% in identifying DoS and DDoS attacks.
• Decision Tree and Multi-layer Perceptron models demonstrated superior performance in detecting DoS/DDoS attacks.
• The developed IDS matched and surpassed current state-of-the-art methods in identifying denial of service attacks.

Conclusions:

• The proposed Machine Learning and Deep Learning-based IDS effectively identifies DoS and DDoS attacks in IoT networks.
• The study highlights the efficacy of Decision Tree and Multi-layer Perceptron models for IoT network security.
• Timestamp analysis and feature selection are crucial for optimizing intrusion detection performance.