Active Queue Management as Quality of Service Enabler for 5G Networks

5G is envisioned as the key technology for guaranteeing low-latency wireless services. Packets will be marked with QoS Flow Indicators (QFI) for different forwarding treatment. 3GPP defines the end-to-end delay limits, but leaves the QoS provisioning methods as implementation dependent. Different services with different constraints will inevitably share queues at some network entity. On the one hand, maintaining the shared queues uncongested will guarantee a rapid packet delivery to the subsequent entity. A brief sojourn time is indispensable for an on time low-latency priority traffic delivery. On the other hand, if shared queues are maintained undersized, throughput will be squandered. In this paper, we propose the use of AQM techniques in 5G networks to guarantee delay limits of QoS flows. Through the evaluation of realistic delay-sensitive and background traffic, we compare different possible solutions. We show that AQM mechanisms together with limited queues, maintain the system uncongested, which reduces drastically the delay, while effectively achieving the maximum possible throughput.