DIDS: A Deep Neural Network based real-time Intrusion detection system for IoT
Monika Vishwakarma, Nishtha Kesswani
Abstract
The number of people using the Internet of Things (IoT) devices has exploded in recent years. The instantaneous development in deploying constrained devices in numerous areas makes them vulnerable to assaults due to limited resources. Advanced cryptography cannot be constructed in these modest battery-powered devices. However, due to the unique properties of the constrained devices, current solutions are insufficient to protect the complete safety scope of IoT networks. An anomaly-based Intrusion Detection System (IDS) is used to identify and categorize assaults. Machine Learning (ML) and Deep Learning (DL) techniques, skilled in embedding intellect in IoT devices and networks, can address various security issues. In this article, we have proposed a deep neural network-based intrusion detection system to identify malicious packets in real-time. We have used newly developed benchmark Netflow-based datasets to train the model. We have proposed a packet capturing and detecting algorithm for real-time attack detection. We also demonstrate the accuracy of our suggested model.