IoT Sensors’ Characteristics used for Behavior Profiling
Alin Zamfiroiu, Radu Marius BONCEA, Ionuţ Petre · 2023
As IoT objects communicate via wireless technologies and are integrated into the larger system of a company or a smart home, attackers who exploit the vulnerabilities of the system can gain easy access to the entire network. Moreover, due to their growth in popularity the average size and complexity of IoT networks are increasing, thus exposing more vulnerabilities. Not only the large number of objects in a network can increase the probability of vulnerabilities, but also their diversity. Being in the same IoT environment, once an attacker gains control of a node, he can perform malicious activities against other devices within the system. For this reason, it is necessary to analyze the behavior of all the sensors and actuators in the system, and, when properties and processes are identified that do not correspond to the classic profile of a sensor, the automatic decision to remove or isolate that module or sensor from the network should be made, until the moment a detailed analysis is conducted. In this research, we identify the indicators and characteristics that allow the realization of a behavior profile for sensors used in IoT systems, but also for larger modules in IoT systems. The research presents their characteristics and how they are measured. Real-time analysis of sensor behavior can identify possible anomalies when attackers try to gain access to certain sensors. By isolating those sensors from the network until they are thoroughly analyzed, the overall cybersecurity of the IoT system is increased.