Speed Tests Questions

[TheLyppardMan]

I'm seeing some strange results when I conduct speed tests. If I run speed tests from my laptop, I get higher download speeds at the BT Wholesale speed test site, but the networx graph upload speed doesn't match what the BT results show (see first screenshot). However, if I run speed tests from my RT-AX88U, the networx graph shows an upload value closer to the result reported by the router (see second screenshot), but slightly less of a match to the (slower) download speed. Both tests were conducted during quiet network times. I'm particularly curious as to why the BT Wholesale test should be showing a much lower download value on the networx graph, even though it is supposed to be monitoring router traffic, rather than just traffic to/from my laptop.

 

[DJones]

I'm seeing some strange results when I conduct speed tests. If I run speed tests from my laptop, I get higher download speeds at the BT Wholesale speed test site, but the networx graph upload speed doesn't match what the BT results show (see first screenshot). However, if I run speed tests from my RT-AX88U, the networx graph shows an upload value closer to the result reported by the router (see second screenshot), but slightly less of a match to the (slower) download speed. Both tests were conducted during quiet network times. I'm particularly curious as to why the BT Wholesale test should be showing a much lower download value on the networx graph, even though it is supposed to be monitoring router traffic, rather than just traffic to/from my laptop.

Your connecting to two different servers located in different areas. Lot of factors play into what could be causing a difference but take it all with a grain of salt.

also the routers speed test is running a test from the source before the other dhcp clients connected to your router meaning highest priority. It’s going to be faster on the router, but it’s still depends on the server your connecting to.

You should be comparing tests with Speedtest.net webbrowser on router vs computer. And speedtest.net CLI on both router and computer. And specify a server if you want the most accurate results.

 

[TheLyppardMan]

I appreciate that different testing sites are going to give different results, but my question really is why the network app running on my laptop gives a reasonably good match with the Ookla test, especially with regard to the reported upload speed, but gives very poor correlation with the BT Wholesale test results. Just to confirm, I've just carried out a speedtest from my laptop at the Ookla web site (so the same conditions as for the BT test, and the results shown on the networx app are much closer to the reported results from Ookla (see additional screenshot below).

 

[DJones]

I appreciate that different testing sites are going to give different results, but my question really is why the network app running on my laptop gives a reasonably good match with the Ookla test, especially with regard to the reported upload speed, but gives very poor correlation with the BT Wholesale test results. Just to confirm, I've just carried out a speedtest from my laptop at the Ookla web site (so the same conditions as for the BT test, and the results shown on the networx app are much closer to the reported results from Ookla (see additional screenshot below).

Indeed a odd discrepancy, since I’m not familiar with networx I can’t really troubleshoot the application itself. But you can see if windows task manager or performance monitor itself sees the same network usage and compare it against Ookla and BT and compare it against what the router sees in network traffic monitor or bandwidth monitor on the router if your router supports it. That would narrow it down to simply being a networx app quirk.

Speedtest operates on TCP idk what BT uses assuming only tcp, but from a bit of googling networx ignores packet overhead and windows broadcast which maybe could cause some discrepancies, but idk why it’s only upload that’s being hit the most in your testing.

 

[TheLyppardMan]

I've just run a couple of BT Wholesale tests from my laptop and Task Manager gives similar results to networx as you can see, so it doesn't appear to be the case that networx is exhibiting some quirky behaviour.

 

[dlandiss]

Another point to consider is that the browser-based Ookla test (speedtest.net) may show lower speeds on many Windows computers than the dedicated Speedtest app. Something to do with the baggage added by Windows, I think.

I get the same speeds (220-230Mbps, paying for 200Mbps) with the little app and my mobile phone, as much as 40Mbps less in a browser on my Win10 laptop.

 

[TheLyppardMan]

I've just run a couple of speedtests using the Ookla Android app and I get similar results to what is shown by the networx app on my laptop but what I am really trying to find out is why the BT Wholesale results don't match. The two screenshots show the difference. The second one is the BT result, 201.13 Mbps download and 73.19 Mbps upload, which as you can see, is not what the networx app is reporting (and as I showed above, the Task Manager in Windows 11 was in broad agreedment with the networx result).

 

[adrenalize]

I've never found the BT Wholesale speedtest particularly consistent - I'd guess they do some smoothing/interpolation to try and give a real world result but it seems to just mess it up quite often. On my BT retail "500/75" it can show just under 500 and about 70 up a lot of the time - but if you look at the Windows task manager/performance a couple of tests previous were only showing ~200Mbps. Heck - it even gave an upload of >90Mbps on one of those which my profile can't even do and certainly wasn't that speed on the task manager. It also seems to show what you have observed - the test is showing 69Mpbs upload, but task manager looks more like 50 Mbps.

I know you are comparing speedtest results to "networx" however I prefer to use a few other speedetst sites - Thinkbroadband, AAISP libre speedtest and openspeedtest. Generally they give me consistent results and I can saturate my line. I'm the only one on the PON at the moment though, so will be interesting to see when other properties upgrade what the contention is like.

I'd just stay clear of the BTW tester and use some different ones!

 

[TheLyppardMan]

I'm going to stop concerning myself with this issue; life is too short. As long as everything is working properly, I'll be happy and the Full Fibre 500 is much better than I had before and that was much better than I had before and that...etc (you get the picture). Hard to believe that when I first started using the Internet, it was via a V90 modem that took absolutely ages to download even a small file and often used to terminate the connection when the download reached 90+% complete (so frustrating). Also, no-one could use the phone when anyone was using the internet, but now, both my VoIP handsets can be in use at the same time.