Lightweight IoT Malware Detection Solution Using CNN Classification
This addresses security challenges for IoT networks, particularly in 5G environments, but is incremental as it applies existing CNN methods to a new domain.
The paper tackles IoT security by developing a lightweight malware detection system using convolutional neural networks (CNN) to recognize malicious behavior in IoT nodes, achieving a solution that can be generalized and applied easily to any network.
Internet of Things (IoT) is becoming more frequently used in more applications as the number of connected devices is in a rapid increase. More connected devices result in bigger challenges in terms of scalability, maintainability and most importantly security especially when it comes to 5G networks. The security aspect of IoT devices is an infant field, which is why it is our focus in this paper. Multiple IoT device manufacturers do not consider securing the devices they produce for different reasons like cost reduction or to avoid using energy-harvesting components. Such potentially malicious devices might be exploited by the adversary to do multiple harmful attacks. Therefore, we developed a system that can recognize malicious behavior of a specific IoT node on the network. Through convolutional neural network and monitoring, we were able to provide malware detection for IoT using a central node that can be installed within the network. The achievement shows how such models can be generalized and applied easily to any network while clearing out any stigma regarding deep learning techniques.