by multiplexing better, you generally need to not differentiate between high and low priority. Being that I've been trying to convince people of this for years

by multiplexing better, you generally need to not differentiate between high and low priority. Being that I've been trying to convince people of this for years - all I can do is try to point at resources that demonstrate how we think about how packets work is frequently wrong, and/or try to provide a demonstrable example. Let's take voip vs netflix over 60ms - netflix, like most DASH video traffic from most video services, is very bursty. It tries to grab a chunk of video, and over the course of 1-2 seconds, fills the pipe, and voip sends one tiny packet every 20ms, which kind of looks like this: What you want: NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN -> you V V V What you get: VVVNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN -> you What AQM does is keep that induced queue short (roughly 16ms with DOCSIS-PIE). The netflix "pulse" goes away and you end up with NNNNNNNNNNNVNNNNNNNNNNNNNN -> you Netflix loads the next segment a tiny fraction slower, but your voip call doesn't get the jitter and latency side-effect anymore.