An Intelligent Two-Phase Automated Architecture for Securing SDN-Based IoT Infrastructure

The Internet of Things (IoT) will bring many opportunities in the next years. However, IoT devices have processing and power limitation. Thus, security remains one of the main challenges. Software-defined networking (SDN) helps traditional IoT infrastructure become manageable and flexible in a centralized fashion. The SDN-IoT architecture tackles the security issue of IoT networks. The proposed architecture adds a new security engine to the controller. The security engine consists of the monitoring, intelligent sub-layer, analyzing/detection engine, reaction, and config engine to automatically monitor, analyze, classify, detect, and generate a proper reaction to the possible threads in two phases. The config engine automatically rearranges the security rules and applies the set of rules as a new configuration to the devices (switches) in the data layer. The intelligent sub-layer uses AI-based feature selection (Bat Algorithm) and classification (Random Forest) algorithms to reveal the possible threats and forward its output to the analyzing/detecting engine to examine it and make the alerts. The cooperation of the intelligent sub-layer and analyzing/detection engine in the two mentioned steps help the system improve the overall system performance and false positive alerts. The proposed architecture follows new security rules based on the network status such as bandwidth minimization and traffic. All the process automatically makes by the security engine and protects the entire network from different threats and attacks.