KDSR: Hybrid Machine-Learning Solution for Intrusion Detection in Fog Computing Environment

Fog computing, that hip cousin of cloud computing, acts as the mediator between cloud data hubs and our gadgets, providing computing, storage, and network services. It is all about putting smarts right at the edge, speeding up those on-the-fly decisions, and cutting down the data traffic to the cloud. Both fog and cloud computing are susceptible to various types of cyber-attacks that could result in unforeseen losses. For instance, malware, phishing, DoS and DDoS, SQL injections, and IoT-based attacks attack has the potential to obstruct verified users. This research introduces an intrusion classification model that harnesses the power of stack machine-learning model by employing a combination of K-Nearest Neighbors, Decision Tree, Support Vector Machine, Random Forest (RF). Comparative analysis with traditional machine-learning approaches highlights the superior performance of our model. The model is intended for deployment in fog layer responsible for monitoring network traffic and detecting cyber-attacks. This implementation ensures the safeguarding of cloud servers and fog layer devices against malicious users, thus ensuring uninterrupted service provision to IoT devices.