A Lightweight Zero-Trust Anomaly Detection Framework for IoT-Enabled 5G/6G Infocommunication Networks

The synthesis of Internet of Things (IoT) devices into present-day infocommunication systems boosted the complexity and liability of 5G and emerging 6G infrastructures. This present paper introduce a new method and caved a lightweight zero-trust anomaly detection framework that associate and collaborate the flow-based network behaviour analysis, a Random Forest classification, and zero-trust decision logic. With the help of this detection component we are able to evaluate using a CICIoT2023 subset containing the following - benign, distributed denial-of-service, and reconnaissance traffic. After several preprocessing of our data from the sources mentioned above, 644,692 records with 39 numerical features were used in total for this paper. And our Experimental results show accuracy of 95.95% , precision of 98.38% , for recall we got 94.06%, and 96.17% F1-score. This Feature importance analysis indicates and tells us that the packet size, average flow characteristics, acknowledgement behaviour, protocol indicators, and timing features are the key contributors to detection. Our results and out put after analying the data demonstrate the potential of combining zero-trust principles with behaviour-based anomaly detection for practical infocommunication security.