Envision, Enhanced, Envisage (EEE) IoT Device Prediction Using Neural Network

Envision, Enhanced, Envisage (EEE) IoT Device Prediction Using Neural Network

연구 분야: Networking

학회: International Research Conference on Computing Technologies for Sustainable Development

Frequently the cause of security issues, Internet-of-things (IoT) devices stand to gain a great deal from automated management. Strong device identification is necessary for this in order to implement the proper network security measures. To tackle this challenge, we investigate methods for precisely identifying IoT devices by observing their network behavior and utilizing strategies that have been suggested by other researchers. A unified architecture for identifying and managing Internet of Things (IoT) devices. Leveraging deep metric representation learning, our approach analyzes network communication data to automatically identify both known and unauthorized IoT devices. We extract relevant features from network traffic data, capturing patterns specific to different device types. These features are used as input for the deep metric learning model. Our model learns a compact and discriminative representation for each device based on the extracted features. By minimizing intra-class variations and maximizing inter-class separations, it enhances device identification accuracy. The architecture defines a unified decision boundary that accommodates diverse IoT devices. This boundary ensures robustness against variations in device behavior and network conditions. In summary, our unified architecture combines domain knowledge with deep learning techniques, providing an effective solution for IoT device identification. Its scalability and adaptability make it suitable for large-scale IoT deployments. The architecture achieves high accuracy, exceeding 99%, in identifying various types of devices, including traffic from smartphones and computers. By automating the device identification process, our solution contributes to better management and security of IoT networks.