RPL-Shield: A Deep Learning GNN-Based Approach for Protecting IoT Networks from RPL Routing Table Falsification Attacks

Considering the growing threats within the domain of the Internet of Things (IoT), this research responds to the pressing need for protection measures that would help to prevent routing table falsification attacks. This research aims at improving IoT networks especially on the area of securing it against such prevalent attacks through a novel deep learning approach. The approach that has been developed is called SecureGuard and it uses the concept of Graph Neural Networks (GNNs) to provide dynamic analysis of evolving network structures. Thus, training SecureGuard on the labeled datasets containing both normal and attack scenarios shows its ability to effectively detect subtle signs of routing table falsification. The achievement of this proficiency sets up a new parameter for identifying, as early as possible, the possible threats on the way. The work provides an exhaustive demonstration of the performance and capabilities of SecureGuard during simulated attacks asserting on its capability of real-time identification and prompt countermeasure against various types of attacks, and proving the model’s high accuracy, precision and recall ratios for accurate distinction of abnormal network traffic and simulated attacks. It is, however, clear that SecureGuard could help to leverage deep learning technologies to improve the robustness of IoT network infrastructures in the face of new complex threats with proven applicability in real-world conditions are still yet to be validated). The results reported here clearly and immediately highlight the necessity of preparing early security measures in the context of IoT related issues, enabling a sound perception of the critical challenges of routing table falsification attacks and their extremely high potential to be detected and reacted to at a very brisk pace.