A comparative study of techniques, datasets and performances for intrusion detection systems in IoT

Abstract

IoT Security is the area concerned with safeguarding connected systems. IoT involves the set-up of various integrated devices. Devices are identified with a unique identifier, and provided with the ability to transfer data over the network opens them up to several serious vulnerabilities, if not appropriately protected. An Intrusion Detection and Prevention System (IDPS) plays a crucial part in discovering and preventing numerous attacks entering the network and provide an uncompromised secure system. The sheer volume of the sensors in a system comes with limitations such as interoperability, scalability, and storage, where security algorithms like IDS couldn’t perform well as it requires a huge amount of labelled data for training, to detect intrusions and ascertain new attacks. Fog computing plays a major role with a decentralized architecture allows IoT devices to compute, make decisions, take actions, and push only relevant information to the cloud. Data availability is closer and can act immediately for the sensitive information, which in turn helps the IDS to perform well using Artificial Intelligence algorithm to detect and prevent various attacks. This paper categorizes the existing recent researches in IoT Intrusion Detection systems using artificial intelligence and fog computing architecture in terms of technical constraints.