Date of Defense
13-6-2025 10:00 AM
Location
E1-1023
Document Type
Thesis Defense
Degree Name
Master of Science in Information Security
College
CIT
Department
Information Systems and Security
First Advisor
Prof. Khaled Shuaib
Keywords
IOTA Tangle, Convergence of IOTA Tangle and AI, Anomaly detection, IOTA Tangle Security, IOTA Tangle and AI security
Abstract
The Internet of Things (IoT) ecosystem has advanced with the advent of Distributed Ledger Technology (DLT) and Artificial Intelligence (AI). Individually, DLT and AI have been explored for enhancement of data management, security, integrity and efficiency of IoT systems. In this thesis, the combined use to apply DLT and AI for network anomaly detection in IoT systems is considered. A framework is proposed to integrate IOTA Tangle, a DLT architecture with Machine Learning (ML)- Random Forest, Decision Trees, and LightGBM, to detect network anomalies in IoT systems. The proposed framework processes network traffic data from UNSW-NB15 dataset and categorizes it into layers based on the OSI model with Smart Contracts for classification. Following this multiple Tip Selection Algorithms (TSAs) for Tangle formations are compared, and the Cache TSA is identified as the best performing with large datasets to form multiple OSI layer specific Tangles. After the formation of the Tangles, ML models are used on it with data mapping to map tangle data to raw dataset and train and test the models on normal vs anomalous traffic. The framework is then evaluated by comparing with traditional ML based anomaly detection on OSI-separated dataset slices without Tangle formation. The results demonstrate that the Tangle structure adds an additional data integrity and security layer with minimal loss of comparability to standard methods. This study contributes to the field of information security by improving IoT data management through IOTA Tangle ensuring integrity of data and Machine Learning to optimize anomaly detection.
INTEGRATING IOTA TANGLE AND ARTIFICIAL INTELLIGENCE (AI) IN IOT NETWORK FOR NETWORK ANOMALY DETECTION
E1-1023
The Internet of Things (IoT) ecosystem has advanced with the advent of Distributed Ledger Technology (DLT) and Artificial Intelligence (AI). Individually, DLT and AI have been explored for enhancement of data management, security, integrity and efficiency of IoT systems. In this thesis, the combined use to apply DLT and AI for network anomaly detection in IoT systems is considered. A framework is proposed to integrate IOTA Tangle, a DLT architecture with Machine Learning (ML)- Random Forest, Decision Trees, and LightGBM, to detect network anomalies in IoT systems. The proposed framework processes network traffic data from UNSW-NB15 dataset and categorizes it into layers based on the OSI model with Smart Contracts for classification. Following this multiple Tip Selection Algorithms (TSAs) for Tangle formations are compared, and the Cache TSA is identified as the best performing with large datasets to form multiple OSI layer specific Tangles. After the formation of the Tangles, ML models are used on it with data mapping to map tangle data to raw dataset and train and test the models on normal vs anomalous traffic. The framework is then evaluated by comparing with traditional ML based anomaly detection on OSI-separated dataset slices without Tangle formation. The results demonstrate that the Tangle structure adds an additional data integrity and security layer with minimal loss of comparability to standard methods. This study contributes to the field of information security by improving IoT data management through IOTA Tangle ensuring integrity of data and Machine Learning to optimize anomaly detection.
