You seem to be using QoS to refer exclusively to stuff like DiffServ. That's wrong, even if it was widely accepted back in whatever decade you were working in n

You seem to be using QoS to refer exclusively to stuff like DiffServ. That's wrong, even if it was widely accepted back in whatever decade you were working in networking. The DiffServ/DSCP mindset for how to achieve QoS is obsolete and inferior, and encourages non-neutral network management to the detriment of everybody except the ISPs that get to charge third parties extra for fast lanes and the few content providers who can afford to pay for the fast lane. If your mental model of QoS treats packet scheduling as a zero-sum game and views latency and bandwidth as inversely related (implied by one of your other comments in this thread), then you're more than a decade out of date. Business models based on such wrong ideas do not deserve any protection from public policy.