The Implementation of a Machine Learning-based Routing Algorithm in a Lab-Scale Testbed

High quality of service (QoS) requires monitoring and controlling parameters such as delay and throughput. Due to network complexity, conventional QoS-improving routing algorithms (RAs) may be impractical. Thus, researchers are developing intelligent RAs, including machine learning (ML)-based algorithms to meet traffic Q oS r equirements. However, most current studies evaluate performance using simulations. Validation requires real-world environment studies, but lab-scale testbed studies are limited. Therefore, we proposed an ML-based RA (ML-RA-t) to improve delay and throughput, evaluated using simulation and a lab-scale testbed. The results show that ML-RA-t predicted the fastest route as compared to RIPv2 routing protocol in simulation and testbed.