Speedtest.net - A Heads Up...

[GabeU]

I had an issue with Glasswire, so I had to uninstall and cleanly reinstall.  Wanting to find out how much data a test at speedtest.net used, I went to their page, cleared my data in Glasswire, and ran the test.  The following was the result, and it shows that a single test used 76MB.  ONE TEST!  If only the problem with testmy.net was no longer a problem.                                                                                                                                                                                

[GW]

You're right Gabe. I noticed that a while back with my phone which has the speedtest app. The speedtest.net auto adjusts file size according to quality of connection.

To provide supporting info, I just ran a few tests. Testing 4G LTE at 102 dBm netted 12.96 down/ 4.3 up Mbps while using 22.5 MB. At 112 dBm getting 2.11 & 0.58 Mbps, only 4.6 MB was used. Switching to wifi on the satellite gave me 30.4 and  6.07 Mbps while consuming 40.3 MB. At 40+ Mbps, 70+ MB per test sounds just about right.

At the primetime G4 speed I used to get and the speeds all the users of badly broken G5 are getting now, it would probably consume about 100 KB

[Amanda]

Hey Gabe

I was testing out speedtest.net and probably flew through 3GB yesterday on our Gen5 I did figure out that when testing 2 different devices I was getting drastically different speeds. Upon checking the server, the devices were picking 2 different servers and when placed on the same ones, produced parallel results. Just something to look out for.

Anyway, our engineers have been testing out 'testmy.net' lately and this morning mentioned that it was looking steady, at least on the Jupiter 2 (E19) units they have set up. I haven't had a chance to play around with it this morning (and I'm pretty sure we're out of data...) but maybe you can give it a go and see if it is still wonky?

~Amanda

[MarkJFine]

@GabeU@Amanda

Noticed on Tuesday that TestMy was more reliable wrt speed consistency as well as latency. Turns out they had been reporting latency in the 200s for quite some time.

Now it's much closer to the 600s, which is probably where it should be.

Edit: Sorry. Wednesday... easy to lose track. Anyway, just checked everything against hnsspeedtest and the Terminal/Gateway Connectivity Test. TestMy and Hughes very similar wrt Down/Up/Ping now.

[GabeU]

@Amanda

@MarkJFine

I, too, noticed the other day that TMN seemed to be working properly and giving more accurate results than it recently had been.  No more of that "choking" I was seeing.  Hopefully it lasts.  

Edit:  Just tried it again a couple of times and it worked the way it should.   I'll try it a few times over the weekend, as well.  

[BirdDog]

I don't know what to believe other than throughput meter and speedtest.net is reporting higher speed. My throughput meter verifies the higher speed. Does seem for me at least on Gen 4 that TMN reports lower than what is actually available speed wise.

Back to back test results:

[MarkJFine]

The thing that bears watching is the amount of data Speedtest uses, as @GabeU noted.

I'd be less tempted to hit that button more than once a day if I knew it was using 3x the data TestMy was.

[BirdDog]

---

@MarkJFine wrote:

The thing that bears watching is the amount of data Speedtest uses, as @GabeU noted.

I'd be less tempted to hit that button more than once a day if I knew it was using 3x the data TestMy was.

---

Not really the point for me. Trying to detrmine the speed available is what I try to detrmine by speed tests. The amount used is really not a factor. Maybe I'm confused once more about what we're talking about.

[BirdDog]

Once more, Speedtest.net is reporting higher speed along with my throughput meter. TMN is reporting lower. I thought we were talking about which is more accurate. At least for me on Gen 4, Speedtest.net seems to be reporting higher verified speed.

[GabeU]

@BirdDog

I think one of the main differences between Speedtest and Testmy is that Speedtest gives a result of what your speed was at the end of the test, whereas TMN gives a result as an average of the whole test.  So, when taking this difference into consideration, they both appear to be accurate when compared to Networx, or at least with how they're giving the results.