A generation of dataset towards an Anomaly-Based Intrusion Detection System to detect Denial of Sleep Attacks in Internet of Things (IoT)
Ishara Dissanayake, Hesiri Weerasinghe, Anuradhi Malshika Welhenge · 2022
In this digital era, Wireless Sensor networks (WSN), or else, Internet of Things (IoT) when these sensor networks are accessible via the internet, is a common and widely used technology. These tiny sensor nodes in these networks help to enhance and automate processes in various fields such as, smart homes and cities, healthcare, smart vehicular networks, Smart Grids and so on. Although this technology is widely used in many fields, these tiny devices are equipped with limited amounts of resources. Limited capacity of battery power is one of the major concerns with IoT when it comes to some specific applications. In addition to this, this limited battery capacity is vulnerable for power depletion attacks which can reduce the lifetime of these devices significantly. So, there is an urgent need of solid solution to preserve this battery power as much as possible. One proper solution for this is, a machine learning based intelligent intrusion detection system that can detect power depletion attacks against these IoT applications. However, the lack of a proper dataset to train and evaluate these machine learning models appears as a major obstacle for an intelligent intrusion detection system. So, in this paper, we are going to discuss how we can use COOJA simulator to generate a proper dataset that can be used to develop and evaluate machine learning models to detect Denial of Sleep attacks against Wireless Sensor Networks. Here, we are going to discuss 3 strategies (UDP flood attack, wormhole attack and externally generated legitimate request attack) to simulate Denial of Sleep attacks for WSNs, so that, any person can generate their own dataset according to their requirements. At the end of the paper, we have evaluated each strategy to deplete the power of the sensor nodes. Then, our conclusions are given.