An Evaluation of Active Queue Management in IoT Network <Abstract> In recent years, IoT technology has been spreading rapidly, with more and more IoT devices have connected to the Internet. In the past, the IoT device was assumingly generated a small amount of data. However, the IoT devices have now supported various types of transmission, such as bulk transfer and streaming. In an IoT network with such a variety of data types, the mixed data flows may not meet the QoS requirement, especially when the router (or network gateway) uses FIFO queue management. The reason is a long flow may put pressure on the queue and increase the queuing delay, which in turn increases the round trip time (RTT). One solution to this problem is to use active queue management (AQM) instead of FIFO queues to detect congestion based on queue length and queuing delay. This function detects congestion quickly and allows the congestion control algorithm running on the sending side to work properly (i.e., adjusting the sending rate and improving the RTT). In this study, we use an emulator to create a topology with contending flows. We then evaluate the performance of six AQM algorithms in terms of throughput and RTT. |