This project will focus on accurate traffic classification in the Internet of Things (IoT). The IoT comprises large numbers of heterogeneous simple devices running single applications, often with little to no security features making them easily compromised and used as tools in cyberattacks. As we become more connected and reliant on the Internet, any form of disruption in connectivity due network anomalies can result in adverse consequences, ranging from loss of productivity and revenue, to destruction of critical infrastructure and loss of life. In the last decade, cyberattacks have increased at an alarming rate, even just based on the reported incidents. We need to be able to classify new traffic types coming from IoT devices accurately and promptly, so that anomalous traffic can be identified and dealt with quickly.
Payload-based (PB) techniques although can reach high accuracy, but suffers from several limitations. The limitations of PB classification are expected to be addressed by statistical-based (SB) techniques. SB approaches are based on flow features and the traffic is classified using Machine Learning algorithms (MLAs). SB classification assumes that specific flow-level features such as flow duration, inter-arrival time, transmitted bytes, packet length and packet size can distinguish different types of traffic flows. We study how unsupervised machine learning can be applied to network anomaly detection in the dynamic IoT environment where previously unencountered traffic types and patterns are regularly emerging and need to be identified and classified. This project involves the study and selection of appropriate MLAs (to be implemented as a proof-of-concept prototype) and identification of those flow features which have the highest impact on the traffic classification accuracy. The success of this project will contribute to making safer cyber-physical systems that are an integral component of the IoT.