IoT-Shield: A Novel DDoS Detection Approach for IoT-Based Devices

The widespread deployment of sensors and linked items contributes to the rising interest in the Internet-of-Things (IoT). These are used in conjunction with other Online services to develop highly sophisticated and profitable cloud-based services. Despite significant attempts to secure them, security management remains a crucial problem for these devices, Because of their intricacy, heterogeneous nature, and resource constraints. This paper presents IoT-Shield, a data mining technique that coincides with a process mining approach for identifying misbehavior in of the kind. IoT-Shield enables the characterization of IoT devices' behavioral models and the detection of possible threats, even in the presence of diverse protocols and platforms. The underlying architecture and components are then described and formalized, and a proof-of-concept prototype is detailed. A real-world traffic dataset known as KDD-NSL is used to evaluate the performance of our technique in a series of comprehensive trials. We conclude that XGBoost is the best available data mining algorithm for the IoT DDoS prediction method. We highlight how consolidating process mining with data mining helps in covering heterogeneity and lowering the computational burden.