The results of the experimental study of the Active Queue Management method at the interfaces of telecommunication networks

Abstract

he paper presents the results of an experimental study of the active queue management method at the interfaces of telecommunication networks. Experimental adequacy and assessment of the effectiveness of the method from the point of view of the main quality of services indicators were carried out at the base of the Cisco Systems laboratory. As a load testing package, the D-ITG package was used to generate and further analyze network traffic. In order to verify the interface bandwidth allocation solutions obtained during the application of the method, they were compared with the WFQ scheduling mechanisms for the average packet delay. To verify the obtained solutions using the active queue management method, they were compared with the AQM WRED mechanisms according to the probability of packet loss. The results of the investigation confirmed the effectiveness of the active queue management method in terms of improving the average packet delay — from 12-17% to 22-25% for high-priority flows (EF, AF41-43) and from 8-12% to 16-19% for low priority flows (AF11-13). Using the active queue management method reduced the probability of packet loss by 7-12% for high-priority flows (EF, AF41-43) and by 10-17% for low-priority flows (AF11-13). It has been established that the recommended area of the practical application of the proposed models for queuing and bandwidth management is the areas of high congestion (more than 80-85%) and congestion of interfaces of telecommunication networks routers, especially in conditions of increased dynamics of their state change.