Detection of IoT Devices That Mine Cryptocurrency

The continuous expansion of the Internet of Things(IoT) market has brought serious security problems. As cryptocurrency attracts more and more people's attention, the price of cryptocurrency has reached unprecedented heights, and now IoT devices are likely to become the target of cybercriminals for stealing computing resources to mine cryptocurrency. This paper proposes a method based on machine learning to detect the existence of malicious miners using IoT devices in a local area network. Compared with previous methods that leverage static signatures or dynamic analysis, this method has low overhead, is easy to maintain, and independent of specific IoT devices and manufacturers. We collected normal traffic from 4 different IoT devices and the traffic of an IoT device that mines the Monero cryptocurrency. Based on the collected data set, 5 machine learning models have been trained to classify normal traffic and mining traffic. Experimental results show that the proposed method effectively detects IoT device mining traffics.