Is Speedtest.net reliable?

raind33r

raind33r

Expert Member
Joined
Oct 13, 2010
Messages
2,664
Reaction score
72
Location
CTN
I'm trying to stream something from the U.K. and it buffers a lot.
Got 4Mbps ADSl line from Telkom

Speedtest.net says download is hunky dory


But http://www.bandwidthplace.com says no way! And I tend to agree with the latter.

bandwidth.jpg
 
HTML:
bandwidthplace readings are way off on my side.

I'm getting way better speeds than indicated by them.


I'm getting 384 speeds according to that site but my downloads perform at 210 k/b
 
OP, do you have uncapped? What are you streaming? Did you try the free Afrihost host account?
 
raind33r said:
I'm trying to stream something from the U.K. and it buffers a lot.
Click to expand...

Sounds more like shaping than anything to do with your line speed.

But yes, Speedtest is a reliable indication of your bandwidth.

I wouldn't trust that 'Bandwidthplace' site one inch.

Superspeed said:
Speedtest.net tests are a farce. All providers cache speedtest.net results. Do a pure download test from a local high speed server or try a lesser known speedtest such as http://capetown.spdtst.saix.net/ to ensure you get true results.
Click to expand...

"Speedtest.net tests are a farce! Don't trust them! Here, use this link to a Speedtest.net server hosted by Telkom instead."
 
It's very reliable, if speedtest does not load, your internet is off :D
 
DA-LION-619 said:
OP, do you have uncapped? What are you streaming? Did you try the free Afrihost host account?
Click to expand...
I've got uncapped yes.
I'm trying to stream live sport. It's impossible to watch with the constant buffering. Hence I think the bandwidthplace is more accurate, as it relates to my streaming experience.
Have not tried Afrihost yet.
 
raind33r said:
I've got uncapped yes.
I'm trying to stream live sport. It's impossible to watch with the constant buffering. Hence I think the bandwidthplace is more accurate, as it relates to my streaming experience.
Have not tried Afrihost yet.
Click to expand...

What's the link to the live sport?

Also, Speedtest's job (and that of any reliable speed test service) is to show your maximum throughput. It doesn't test for shaping, throttling or any similar issues.
 
raind33r said:
You won't be able to watch without a subscription ;) http://www.psasquashtv.com
Click to expand...

Well, I signed up for a free account and its videos (non-live) stream at around 188kB/s, which is obviously more than enough for a 4Mbps line. My money remains firmly fixed on your ISP shaping the protocols.

Nod said:
Iperf will give more accurate results. Speedtest.net only uses 3 simultaneous connections, IIRC.
You could also try http://speedof.me/ ...
Click to expand...

Funny thing about most of these speed tests is that they suck.

Let me try to illustrate using my line. My bandwidth is 600Mbps down and 100Mbps up.

Speedof.me:

140515163934-38006.png


Note that it also says it uses a Stockholm server.

BandwidthPlace:

5375c048b15db1.78390895-15277928.png


It claims to have used the UK server since it's "the fastest." The Netherlands and German server fared far worse, despite being closer.

As for Iperf... I don't have the willpower in me to try and figure that out and the "http://dast.nlanr.net/Projects/Iperf/" URL is dead.

Here are the results of a local speed test provider called Bredbandskollen:

Wp8SqF3.png


Here are the results from Speedtest.net to... let's say London again for teh lulz:

3504018231.png


And Speedtest.net to "recommended":

3504020529.png


The only two that reflect my actual bandwidth... You guessed it, Speedtest and Bredbandskollen.

Also... Jeez, BandwidthPlace and Speedof.me need to realise FIVEBAJILLIONMEGAPIXEL images don't make for easy viewing... Especially if all they're illustrating are three simple numbers (up, down and ping).
 
Last edited:
Lycanthrope said:
Well, I signed up for a free account and its videos (non-live) stream at around 188kB/s, which is obviously more than enough for a 4Mbps line. My money remains firmly fixed on your ISP shaping the protocols.



Funny thing about most of these speed tests is that they suck.

Let me try to illustrate using my line. My bandwidth is 600Mbps down and 100Mbps up.

Speedof.me:

140515163934-38006.png


Note that it also says it uses a Stockholm server.

BandwidthPlace:

5375c048b15db1.78390895-15277928.png


It claims to have used the UK server since it's "the fastest." The Netherlands and German server fared far worse, despite being closer.

As for Iperf... I don't have the willpower in me to try and figure that out and the "http://dast.nlanr.net/Projects/Iperf/" URL is dead.

Here are the results of a local speed test provider called Bredbandskollen:

Wp8SqF3.png


Here are the results from Speedtest.net to... let's say London again for teh lulz:

3504018231.png


And Speedtest.net to "recommended":

3504020529.png


The only two that reflect my actual bandwidth... You guessed it, Speedtest and Bredbandskollen.

Also... Jeez, BandwidthPlace and Speedof.me need to realise FIVEBAJILLIONMEGAPIXEL images don't make for easy viewing... Especially if all they're illustrating are three simple numbers (up, down and ping).
Click to expand...

Oh stop bragging with your speed :p
Darn 1st world countries :whistle:

Maybe someday we'll get there.
 
raind33r said:
Oh stop bragging with your speed :p
Darn 1st world countries :whistle:

Maybe someday we'll get there.
Click to expand...

Eish... Did you have to quote the pics too? :p

I'm not bragging though, I'm just trying to illustrate my point: Speed tests serve only one real purpose and that's to test your maximum throughput. They don't test which protocols are shaped or whether or not your line is throttled (well, unless your bandwidth as a whole is throttled by your ISP).

And the best one there is, is Speedtest.net.

Although, even there, it's only as accurate as the server you use to test with, since they are almost all entirely independently hosted and can be prone to problems.
 
Lycanthrope said:
As for Iperf... I don't have the willpower in me to try and figure that out and the "http://dast.nlanr.net/Projects/Iperf/" URL is dead.
Click to expand...

iperf is pretty easy. Working iperf servers are listed on wikipedia.
My results, using 10 connections:
Code: 
$ iperf.exe -c iperf.eltel.net -P 10 -m -t 10
------------------------------------------------------------
Client connecting to iperf.eltel.net, TCP port 5001
TCP window size: 64.0 KByte (default)
------------------------------------------------------------
[ 12] local 10.0.180.59 port 8556 connected with 217.170.67.78 port 5001
[  8] local 10.0.180.59 port 8552 connected with 217.170.67.78 port 5001
[ 11] local 10.0.180.59 port 8555 connected with 217.170.67.78 port 5001
[  5] local 10.0.180.59 port 8549 connected with 217.170.67.78 port 5001
[  7] local 10.0.180.59 port 8551 connected with 217.170.67.78 port 5001
[  3] local 10.0.180.59 port 8547 connected with 217.170.67.78 port 5001
[ 10] local 10.0.180.59 port 8554 connected with 217.170.67.78 port 5001
[  9] local 10.0.180.59 port 8553 connected with 217.170.67.78 port 5001
[  6] local 10.0.180.59 port 8550 connected with 217.170.67.78 port 5001
[  4] local 10.0.180.59 port 8548 connected with 217.170.67.78 port 5001
[ ID] Interval       Transfer     Bandwidth
[ 10]  0.0-10.0 sec    744 KBytes    609 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[ 12]  0.0-10.0 sec  1.40 MBytes  1.17 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  5]  0.0-10.1 sec  1.79 MBytes  1.49 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.1 sec  1.34 MBytes  1.11 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  4]  0.0-10.1 sec    600 KBytes    488 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  9]  0.0-10.1 sec  1.28 MBytes  1.07 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  6]  0.0-10.1 sec    632 KBytes    513 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  8]  0.0-10.1 sec    560 KBytes    454 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  7]  0.0-10.1 sec  1.86 MBytes  1.54 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[ 11]  0.0-10.2 sec    792 KBytes    636 Kbits/sec
[SUM]  0.0-10.2 sec  10.9 MBytes  8.97 Mbits/sec
 
Nod said:
iperf is pretty easy. Working iperf servers are listed on wikipedia.
My results, using 10 connections:
Code: 
$ iperf.exe -c iperf.eltel.net -P 10 -m -t 10
------------------------------------------------------------
Client connecting to iperf.eltel.net, TCP port 5001
TCP window size: 64.0 KByte (default)
------------------------------------------------------------
[ 12] local 10.0.180.59 port 8556 connected with 217.170.67.78 port 5001
[  8] local 10.0.180.59 port 8552 connected with 217.170.67.78 port 5001
[ 11] local 10.0.180.59 port 8555 connected with 217.170.67.78 port 5001
[  5] local 10.0.180.59 port 8549 connected with 217.170.67.78 port 5001
[  7] local 10.0.180.59 port 8551 connected with 217.170.67.78 port 5001
[  3] local 10.0.180.59 port 8547 connected with 217.170.67.78 port 5001
[ 10] local 10.0.180.59 port 8554 connected with 217.170.67.78 port 5001
[  9] local 10.0.180.59 port 8553 connected with 217.170.67.78 port 5001
[  6] local 10.0.180.59 port 8550 connected with 217.170.67.78 port 5001
[  4] local 10.0.180.59 port 8548 connected with 217.170.67.78 port 5001
[ ID] Interval       Transfer     Bandwidth
[ 10]  0.0-10.0 sec    744 KBytes    609 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[ 12]  0.0-10.0 sec  1.40 MBytes  1.17 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  5]  0.0-10.1 sec  1.79 MBytes  1.49 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0-10.1 sec  1.34 MBytes  1.11 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  4]  0.0-10.1 sec    600 KBytes    488 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  9]  0.0-10.1 sec  1.28 MBytes  1.07 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[  6]  0.0-10.1 sec    632 KBytes    513 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  8]  0.0-10.1 sec    560 KBytes    454 Kbits/sec
[ ID] Interval       Transfer     Bandwidth
[  7]  0.0-10.1 sec  1.86 MBytes  1.54 Mbits/sec
[ ID] Interval       Transfer     Bandwidth
[ 11]  0.0-10.2 sec    792 KBytes    636 Kbits/sec
[SUM]  0.0-10.2 sec  10.9 MBytes  8.97 Mbits/sec
Click to expand...

Code: 
iperf.exe -c iperf.eltel.net -P 1
0 -m -t 10
------------------------------------------------------------
Client connecting to iperf.eltel.net, TCP port 5001
TCP window size: 64.0 KByte (default)
------------------------------------------------------------
[ 11] local 192.168.1.2 port 63564 connected with 217.170.67.78 port 5001
[ 12] local 192.168.1.2 port 63565 connected with 217.170.67.78 port 5001
[  9] local 192.168.1.2 port 63562 connected with 217.170.67.78 port 5001
[  8] local 192.168.1.2 port 63561 connected with 217.170.67.78 port 5001
[  7] local 192.168.1.2 port 63560 connected with 217.170.67.78 port 5001
[ 10] local 192.168.1.2 port 63563 connected with 217.170.67.78 port 5001
[  6] local 192.168.1.2 port 63559 connected with 217.170.67.78 port 5001
[  4] local 192.168.1.2 port 63557 connected with 217.170.67.78 port 5001
[  3] local 192.168.1.2 port 63556 connected with 217.170.67.78 port 5001
[  5] local 192.168.1.2 port 63558 connected with 217.170.67.78 port 5001
[ ID] Interval       Transfer     Bandwidth
[  6]  0.0-10.1 sec  10.0 MBytes  8.33 Mbits/sec
[  6] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  8]  0.0-10.1 sec  13.8 MBytes  11.4 Mbits/sec
[  8] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  5]  0.0-10.1 sec  12.1 MBytes  10.1 Mbits/sec
[  5] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[ 12]  0.0-10.2 sec  6.25 MBytes  5.15 Mbits/sec
[ 12] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  7]  0.0-10.2 sec  9.00 MBytes  7.40 Mbits/sec
[  7] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  9]  0.0-10.2 sec  10.9 MBytes  8.91 Mbits/sec
[  9] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[ 11]  0.0-10.3 sec  8.25 MBytes  6.71 Mbits/sec
[ 11] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[ 10]  0.0-10.3 sec  8.12 MBytes  6.60 Mbits/sec
[ 10] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  3]  0.0-10.4 sec  6.25 MBytes  5.06 Mbits/sec
[  3] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[  4]  0.0-10.8 sec  9.00 MBytes  7.00 Mbits/sec
[  4] MSS and MTU size unknown (TCP_MAXSEG not supported by OS?)
[SUM]  0.0-10.8 sec  93.6 MBytes  72.8 Mbits/sec

Thanks :)
 
Superspeed said:
Speedtest.net tests are a farce. All providers cache speedtest.net results. Do a pure download test from a local high speed server or try a lesser known speedtest such as http://capetown.spdtst.saix.net/ to ensure you get true results.
Click to expand...

I have tried them and sometimes get impossible speeds for my area. Like this morning 95,98Mbps download speed.I can not even achieve such speeds in my dreams.
 
Speedtest is the real deal, the others are all bullsh*t, I dont even trust the myBB one.
 
Rickster said:
Speedtest is the real deal, the others are all bullsh*t, I dont even trust the myBB one.
Click to expand...

The MyBB one is using Ookla, aka Speedtest, in the backend. It simply have predefined servers to test from.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X