AQM technologies were described as necessary best current practice by the IETF back in 1992, and recommendations revised in 2013 https://datatracker.ietf.org/do

AQM technologies were described as necessary best current practice by the IETF back in 1992, and recommendations revised in 2013 https://datatracker.ietf.org/doc/rfc7567/ https://datatracker.ietf.org/doc/rfc7567/ because the algorithm for it chosen in 1992 (RED) didn't work well enough. Two new algorithms appeared in 2012 (Pie and codel/fq-codel) that did. I was very frustrated by the network neutrality "debate", partisans excluding the idea that it was actually a huge technical flaw with how the internet is structured that was at the root of the problem - when many flows attempted to co-exist: https://blog.cerowrt.org/post/net_neutrality_customers/ https://blog.cerowrt.org/post/net_neutrality_customers/ Now Solved. Thoroughly. By those two theoretical breakthrus. Since, zillions of knowlegable users and new products (like those from openwrt, eero, google wifi, and many others) managed to fix the underlying bufferbloat problem for themselves via "smart queue management" or "optimizing for conferencing and gaming", but it required manual configuration and tweaking and the right place for better bandwidth does indeed lie within the ISPs shapers and CPE. And on by default. However, it's not as simple of "call of duty or zoom" going "ahead". Better multiplexing (FQ), and shorter queues (AQM) lead to those applications' packets mixing into the heavier netflix flows without manual intervention. Comcast