FF dev's did admit on a bug report I made some time ago they use an old network API on windows, it mostly affects upload speeds, but I guess could also potentially affect download as well.

In reply to a post by jchamier:

I wonder where it read and ingested that... I've seen plenty of Windows workstations and servers run (on symmetric LANs) without trouble. I worry ChatGPT and similar just suck up forums (like this) with no actual proof.

It does. It's why I flagged it clearly as AI generated. Worth noting a Linux user may not have AV and Windows more likely to.. this may taint 'Windows' problems more.. but equally it's possible that some combinations of issues like specific drivers/chipsets causing issues. We've had fun on servers with Intel cards working nicely in Linux but Broadcom not (10GE SFP cards--I forget the example but something like that)... quirks like this do exist and the average Windows user would be less knowledgable to avoid or fix them.

Also a test on a single Layer-2 LAN may differ from a transfer over Internet with different settings/MTUs, etc..

seb

In reply to a post by Taras:

just for fun.

linux, with 10gbit nic, to 10gbit switch then to 10gbit port on router, Router to ont is 2.5gb (since the ont is 2.5gb) both via ipv6

https://www.thinkbroadband.com/speedtest/17683199384...

windows with 10gbit nic to 10gbit and then same as the above
https://www.thinkbroadband.com/speedtest/17683202245...

interestingly the server has a lower spec'd cpu 😁 but that shouldn't matter

Is your service 1Gbps?

Just for fun. Installed Windows 7 Ultimate x64 with patches security updates with firefox running Intel i7 Sandybridge PC using 1Gbps ethernet on the pc.

https://i.ibb.co/nsKvXJMd/W7.jpg

https://i.ibb.co/KzNyRgGw/W7-updates.jpg

https://i.ibb.co/tTwGqYZD/W7-Device-Manager.jpg

https://i.ibb.co/Tqt4J44L/W7-Firefox.jpg

speedtest using Windows 7 https://www.thinkbroadband.com/speedtest/17684011852...

In reply to a post by adslmax:

speedtest using Windows 7 https://www.thinkbroadband.com/speedtest/17684011852...

Looks pretty good.. 939.3Mbps multithreaded average burst to 955.8Mbps - that's pushing the limits of what is possible on 1Gbps.

In reply to a post by seb:

Is your service 1Gbps?

yes, 1000/115 on a 2.5gbit ont. I felt those results where bad. I think theres a couple of heavy users on the pon i'm on. Is there extra padding on ipv6 vs v4 (not really ever looked into it)?

Your BQM still isn't looking great for FTTP, would be good if you could get to the root cause of those spikes. May be worth considering overlaying the BQM graph for several days to see if there is a pattern to it.

In reply to a post by seb:

We've had fun on servers with Intel cards working nicely in Linux but Broadcom not (10GE SFP cards--I forget the example but something like that)... quirks like this do exist and the average Windows user would be less knowledgable to avoid or fix them.

Yes, drivers and NIC vendors have been perpetual hassles; and why trying to run vendor certified drivers and stacks is the only way on a proper data centre server setup. Once you are 10GbE or faster then it goes wrong fast!

Also a test on a single Layer-2 LAN may differ from a transfer over Internet with different settings/MTUs, etc..

True, internet is always much more complex, but I was thinking across various WAN links, but all inside the UK.

Its been a while for me though... its all now cloud :-/

Cloud isn't really any different to anything else on the Internet.

In reply to a post by PCJM40:

Your BQM still isn't looking great for FTTP, would be good if you could get to the root cause of those spikes. May be worth considering overlaying the BQM graph for several days to see if there is a pattern to it.

Don't think can't do anything about it. The same spikes with my previous FTTC and G.fast and now with FTTP.

Show me your current live BQM?