Smart Web Attack Detection in IoT Using Multi-Model

Abstract

The objective of this research paper is the soaring increase of the. Internet of Things (IoT) has posed serious security threats. on our critical infrastructure systems. Devices with limited resources have a tendency of transmitting information without any sending. properly checking the payload. The present paper presents a. new Zero-Trust IoT Broker Architecture that uses Ensemble The Deep Learning methods, such as CNN, LSTM, and. MRN, along with AES-128 encryption, which is secure adequate to use with blockchain. The Zero-Trust Architecture uses an early-fusion hybrid "Max-Confidence" Ensemble. Network on ingestion layer to detect and remove. harmful payloads, such as XSS, SQL, and Path Traversal attacks, until they arrive at the central repository of clouds. The ensemble network has superior accuracy to. When tested on, the results of F1-score were compared to other models. various attack datasets. The encrypted storage layer guarantees that information stored is consistent post- ingestion