Detection of Distributed Denial of Service in the Application Layer of Iot Using Machine Learning

Cybersecurity prevents harmful assaults on computers, networks, and mobile devices. Included in the category of cyber-attacks are threats like as malware, distributed denial of service assaults, phishing, spoofing, code injection, and supply chain attacks. Massive denial-of-service assaults are destructive. As part of a Distributed Denial-of-Service (DDoS) assault, hackers bypass botnets and go straight for the server. In order to temporarily block or suspend a website, these botnets send a flood of requests to the server. The server typically takes requests and returns responses. There will be a brief outage if the server gets more requests than it can handle. While this attack makes the service inaccessible, it prevents data theft. A machine-learning approach is used to detect server-side DDoS assaults. The term "machine learning" refers to software that can learn new tasks and solve old ones without any input from a human. Learning Vector Quantization, Support Vector Machine, Naive Bayes, K-Nearest Neighbour, Random Forest, and Decision Tree are the six ML classification models used since these ML procedures enhance accuracy. The main goal of this research is to use the CIDAD dataset to find ways to protect the application layer of the Internet of Things (IoT) against Distributed Denial of Service (DDoS) attacks. The lower layers of the IoT network architecture are disregarded by this dataset, which comprises information at the COAP level.