IPAssess: A Protocol-Based Fingerprinting Model for Device Identification in the IoT

The Internet of Things (IoT) has become a widely prevalent concept as it has dramatically advanced the ability to communicate and exchange data between various connected devices. With its success and growing need, many threats and attacks against IoT devices and services have exponentially increased. An increase in knowledge of IoT-related threats and adequate monitoring technologies have helped develop the potential to detect the threats. There have been various studies on fingerprinting based approaches on device identification but none have taken into account the full protocol spectrum. IPAssess is a novel fingerprinting based model which takes a feature set based on the correlation between the device characteristics and the protocols and then applies various machine learning algorithms: Random Forest, Decision Tree, K-Nearest Neighbour (KNN), Naive Bayes, and Gradient Boost (XGB), to perform device identification and classification. We have used aggregation and augmentation to enhance the algorithm. In our experimental study, IPAssess performs IoT device identification with a 99.6% classification accuracy.