Mitigating Security Threats in IoT Networks Using Big Data Analytics and On-Device Modeling

The rapid increase of IoT devices created the modern digital infrastructures but it has also added crucial security challenges due to the scale and heterogeneity of IOT devices. This paper presents a collaborative security framework model that uses big-data analytics & Hybrid on-device modelling that allows us to address the security threats appearing in an IoT ecosystem. The framework uses big data analytics to process huge amounts of IoT traffic data in real time, recognizing patterns, detect possible threats and creates a miniature model that can be deployed on IOT device. On-device modeling helps to ensure that threats can be handled right there on-device, limiting cloud-based infra and latency dependencies. The framework also focus on using an AHD Model (Anomaly Hash-out Delta) using device behavior profile for permissible actions, which helps anomaly handling in a very lightweight machine learning model.