Dynamic Adaptation Attack Detection Model for a Distributed Multi-Access Edge Computing Smart City.

Area of Science:

• Cybersecurity
• Network Security
• Internet of Things (IoT)

Background:

• Internet of Things (IoT) networks are vulnerable to attacks that compromise network integrity and resources.
• Smart cities rely on IoT for various services, making their security paramount.

Purpose of the Study:

• To propose an intelligent automation detection model (IADM) for detecting and preventing malicious network traffic in IoT-based smart cities.
• To enhance security in distributed multi-access edge computing environments within smart cities.

Main Methods:

• The IADM utilizes a two-phase approach: intelligent process automation (IPA) for initial detection and reinforcement learning for dynamic adaptation.
• Phase one involves modules for dataset collection, detection, analysis, rule definition, and database management, using classifiers like Random Forest Trees (RFT), k-Nearest Neighbor (K-NN), J48, AdaBoost, and Bagging.
• Phase two employs reinforcement one-shot learning for adapting to new threats and detecting zero-day attacks.

Main Results:

• The proposed model achieved high accuracy rates (approx. 98.8%) using RFT, K-NN, and AdaBoost classifiers.
• Classifiers like AdaBoost and Bagging demonstrated strong performance, with AdaBoost achieving 98.9% accuracy when based on J48.
• The model exhibited low error rates and high precision, recall, and F1-measure scores, indicating robust detection capabilities.

Conclusions:

• The IADM effectively detects and prevents malicious activities in IoT networks within smart cities.
• The auto-adaptive nature of the model, using reinforcement learning, allows for rapid detection of evolving threats and reduces false positives.
• The IADM improves intrusion detection system (IDS) performance by optimizing memory and time consumption, crucial for resource-constrained IoT devices.