Detecting Brute-Force Attack in IoT Device using Network Flow Data
Tobechukwu Treasure Osueke · NORMA · 2018
The concept of Internet of Things has materialize the vision of a society where users, computing systems and everyday objects possessing sensing and actuating capabilities interact in a manner that is convenient and economical using the Internet. Advancement in IoT have led to the adaptation of current Internet architecture and IP-based communication protocols. Such communication protocols have been deployed with components having low power, low-storage and constraints resources. With the massive increase in number of IoT devices accross the glob and the rapid development of enabling communication protocols, security should be considered. Brute-force attack is one of the most prevalent attacks that targets Internet enabled devices. IoT objects have suffered from this attack type, leading to disruption of services, data confidentiality and integrity. The goal of this work is to improve the security of IoT devices by building and evaluating five models to detect brute force attack on telnet and secure shell (SSH) communication protocols respectively. The models were implemented using five different machine learning classification algorithm and then evaluated to determine which of the algorithms is more efficient and adequate. Our approach uses network flow data and puts into consideration the fact that IoT devices are composed of constraints resources.