Modelling a Novel Linear Transformed Attention Mechanism for Intrusion Detection Using Learning Approach
R Vijay Anandh, Vishakha Sangam Rane, Sangam Subhash Rane, S. Vijayakumar, P. Tamilarasan, Namratha Gopinath · 2024
In the current global scenario of Internet of Things (IoT) pervasive development is inevitably seen. The smart networks faced cyber-attack in critical fault-lines. This situation called for IoT security imminently. The security breach not only affected the IoT but the smart devices which are deployed as botnet connected to internet across the globe is also brought into danger, this in a large view makes the whole Internet ecosystem exploited. Internet became paralyzed when the video surveillance devices were compromised by malware through an attack called denial of service (DoS). In both terms like diversity and complexity pervasive growth of security attack can be seen in the recent past. So, it is important to formulate a technique to analyse the context of IoT to detect and prevent attacks. By evaluation of defence techniques that are present, the presented survey can classify the challenges and threats for IoT networks. Network Intrusion Detection Systems (NIDS) is focused in our study. The presented LTMA model reviews open-source and free network sniffing software as well as datasets and NIDS implementation tools which already exists. Finally, our system performs the surveys and then analyses the same and compares and contrasts the contemporary NIDS in terms of methodologies used for detection, its architecture, strategies employed for validation, deployments of algorithm and threats treated in IoT context. The review of the study in (ML) machine learning and traditional NIDS techniques are discussed and scope for future directions are presented.