Web based speedtest is not accurate?

[DCG909 0]

I'm running a trail to see what my speed would be if I were to tunnel all my traffic trough the VPN (router sided).

 

I first started of with the web based speed test, which would yield me a 30/30 download and a 120/NA upload (out tunnel/in tunnel respectively).

I thought these values were a bit low/off, since I've got a 250 Mbit/s fiber connection.

Then I tried to download some games trough steam (I checked the client application to verify my data was being sent trough the tunnel), which yielded me a tad slower download than I was used to, 31.4 MB/s instead of 32.8 MB/s (I also just did a speedtest trough speedtest.net, to verify the result, after reading a topic of steam bypassing the VPN, which confirmed the same data as trough the steam download).

 

Afterwards I did another web based test with the buffers to detect (had it at 100MB), and this yielded the following results:

 

Down: 30.537 Mbit/s Out, 30.470 Mbit/s In (99%), 20MB - Up: 114.628 Mbit/s Out, 115.876 Mbit/s In (101%), 20MB - Date: Sun, 05 Jun 2016 09:43:19 GMT - Buffers: 20MB/20MB - Laps: 3, Time: 25.29 secs
 

So there might be a little bug somewhere, or I'm interpreting the data wrong

[Staff 9781]

Hello,

 

yes, it can be inaccurate, just like any other speed test. It's very hard to have an accurate web based speed test. Our speed test is useful in determining an approximate out-tunnel/in-tunnel performance ratio and make comparisons. Do not rely on it too much for absolute values. In your case, the outcome is excellent. Even considering the potential error, the performance inside the tunnel appears almost equal to the performance outside the tunnel, which is an excellent outcome.

 

Kind regards

[go558a83nk 352]

First thing, have you tested several different servers?  Getting only 30mbit/s out of the tunnel when your ISP line is 250mbit/s is losing a lot.  Maybe some other servers will have a better route to you.

 

After that, be sure to try different ports and protocols and also try SSL tunnel.  I regularly see people around 200mbit/s in the status->top users page, so I know it's possible.

[DCG909 0]

@go558a83nk.

 

I chose a couple of servers/methods of which I know they can fill my bandwith to the max (if they couldn't, I would have chosen different ones), you can't however chose a server on the web based Air VPN speedtest

I made this OP to tell there was a great difference between the web based test and real world performance, take a peek at Hassaleh's usage today, specifically after 08:00.

 

@Staff.

That kind of test senario (only comparison between with and without tunnel) did cross my mind, thanks for confirming this

I have to say, I expected a much much bigger difference between with and without tunnel, so I am most pleasantly surprised

[go558a83nk 352]

@go558a83nk.

 

I chose a couple of servers/methods of which I know they can fill my bandwith to the max (if they couldn't, I would have chosen different ones), you can't however chose a server on the web based Air VPN speedtest

I made this OP to tell there was a great difference between the web based test and real world performance, take a peek at Hassaleh's usage today, specifically after 08:00.

 

@Staff.

That kind of test senario (only comparison between with and without tunnel) did cross my mind, thanks for confirming this

I have to say, I expected a much much bigger difference between with and without tunnel, so I am most pleasantly surprised

 

oh, I understand now. I didn't read carefully enough the first time.  I didn't see MB/s.