Towards Development of Memory-Efficient Intrusion Detection System Against RPL Attacks

Due to vulnerability of IPv6 Routing Protocol for Low Power Lossy Networks (RPL) to various resource-targeting attacks, there is ongoing effort to develop intrusion detection programs. However, most of these algorithms prioritize detection accuracy over efficiency. This study aims to address both detection performance and memory requirements in developing an intrusion detection system against blackhole, DAG inconsistency, and decreased rank attacks. To this end, non-dominated sorting genetic algorithm and strength pareto evolutionary algorithm, commonly used multi-objective approaches in the literature, are integrated into the genetic programming to decrease read-only memory (ROM), random access memory (RAM) consumption while enhancing detection accuracy. The results demonstrate a significant reduction in memory consumption while maintaining, or even improving, detection accuracy.