Speedtest node error

General
21

Ah, this thread is now really getting interesting.
To start, I agree with your opinion on NR. Taking NR plus MQ, I can "taste" where it originated... actually, I was part of that (Big Blue) team at the time... also did quite a bit of development on VM internals. Those were the days...

On measurements: DSLreports is an online measurement, not suited for my purpose. I have two different networks. One is at home, where my provider delivers a meager 21 Mb/s via copper wire. My house is simply a bit to far from the backbone switch, I guess. But it is sufficient for now.
My second net is in my holiday home, even more remote. Via copper I get about 4 to 6 Mb/s, not quite enough. So they gave me a hybrid modem: partly runs on copper, the second modem is attached to the first with ethernet cable and provides additional bandwidth from 4G mobile. The main modem combines this, and delivers between 30 and 45 Mb/s.

I wanted to have a simple tool that I can run on both subnets, and collect the data centrally. A couple of R-Pi's with Domoticz, MQTT, Node Red and speedtest...

I've also had a SamKnows box running in my home. Probably a reliable way of measuring, but somehow the reports are not what I was looking for. Not sure if they have an API to plug into and pull the data.

22

You tickled my curiosity, so I ran a test on four platforms. Note that they were run in sequence, so a slight difference is to be expected
DSLreports: down 15,98, up 3,85
speedtest.nl: down 16,6, up 4,64
Ookla online speedtest: down 16,0, up 4,14
R-Pi with Ookla: down 15,89, up 4,28

FIrst three were run on a laptop, connected via WiFi. Given the internet speeds that my provider offers, WIFi is not really the bottleneck here.
Lasr measurement is run on a Raspberry Pi 4B with 8GB, ethernet cable connected.

Down speeds are lower that what I mentioned, since my wife is watching a show on TV. I tried to convince her to switch it off for me, but I lost that argument.... :slight_smile:

23

Those are well within expected variations. Also, don't forget that speedtests are highly dependent on the server at the other end. When using Ookla, I always try to use the same server and it certainly isn't one close to me. If I chose anything even remotely close, the results are always much lower, presumably because the server doesn't have sufficient capacity.

Also 16MB is around 80% of your stated 21MB circuit which is probably about right anyway.

24

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.