cloneman wrote: > I don't know if the Windows 10 updates multithreading problem is still > relevant for many users, but it certainly still affects me as user with > low bufferbloat, low latency, and only moderate bandwidth (50mbit , 4ms > idle, ~9ms loaded) > In any case, I have submitted official feedback to microsoft, as I've > exhausted any possible workarounds on my end short of implementing a > Windows update cache server on my LAN (go figure, apparently the cache > servers download with a small number of threads). I guess either because the person implementing the cache system didn't work as hard, or because they recognized they needed to do this in the "background" > After posting this link -- I'm done advocating for this issue. I think > valve steam has made some improvements, they still use many threads, but > somehow, it doesn't create as many issues.There seems to be no interest in > most discussion forums to explore this with any depth -- or even admit that > 20 connections to 1-2 servers is problematic. This is where some of the desire to have a "worse than best effort" queue came from :-) > Feedback hub link > https://aka.ms/AA7zg1r Not a public link, I think. What's sad is that Windows 2000 was the only implementation of the diffedge that we needed to make end-system driven QoS work, and then they abandonned it. -- Michael Richardson , Sandelman Software Works -= IPv6 IoT consulting =-