Data Clustering for Categorizing Normal and Unusual IoT Network Traffic to Identify Attacks

Abstract

Internet of Things (IoT) devices are becoming increasingly prevalent as time goes on, as they present a means of connectivity that is both straightforward and efficient for the end user. These devices are being used everywhere, like in homes, businesses, and people's pockets. Because of the amount of connectivity that they allow, network traffic security is a growing concern, especially as the devices become used in more sensitive environments. Data clustering, a machine learning technique used to group data points, is one solution to aid in classifying network data. By organizing network traffic as normal or unusual, direct and indirect attacks can be identified. This paper will compare the use of various data clustering algorithms to aid in analyzing IoT network traffic and determine if an attack was attempted or not.