The World of Statistics in the Service of the Smart Homes Security
Internet of Things (IoT) paradigm has become extremely popular the last decade, since it offers the ability to create a network infrastructure of objects that follow different wireless network protocols, such as the Bluetooth, the ZigBee and the Z-Wave. Therefore, since the popularity of the IoT is getting more considerable, the need of shielding the IoT security is becoming more urgent and indispensable. Could the world of statistics enhance the security of the Smart Homes?
In GHOST solution, statistics and machine learning techniques are combined to strengthen the security of the IoT ecosystem as described next. Each IoT device communicates with the gateway, this communication between IoT devices and gateway is recorded in GHOST solution and the related data are very useful for advanced statistical and machine learning methodologies. Hence, using statistical measures related to the number and to the size of the communication packets, along with conducting sophisticated statistical procedures such as the Principal Components Analysis, a network behavior can be extracted for each IoT device. Next, the IoT devices could be grouped into behavioral clusters/templates with the Density-based spatial clustering of applications with noise, according to their network traffic and these templates could be used to detect possible anomalies.
The aforementioned procedure constitutes the training phase of this methodology, whereas the monitoring phase is presented as follows. For each IoT device, the current network traffic (in short-term and long-term intervals) is compared with the network traffic of its template and if there is statistical significant difference in the behavior, as depicted by the Euclidean distance from the template center, then a possible threat is detected. The larger the distance from the center, the more possible to detect a true attack and not to have a false alarm by the algorithm.
Emerging IoT technology, undoubtedly, requires innovative methodologies in order to protect the security of the smart homes that consist of many and diverse IoT devices. For this reason, GHOST solution exploits the world of statistics that offers valuable tools, in order to shield the IoT ecosystem security and to ensure smart home users about their privacy and safety.