Telkom Internet Capped/Uncapped User Feedback (Pt2)

bwana said:
I haven't don't alternate speed tests. My downloads tend to run at full 4mbps speeds so I'm not complaining as such - I'm getting what I'm paying for.

I'm a little concerned Telkom might be getting more credit than it is due when it comes to MyBB's speedtest results though. This may well be a flaw in the design of the speedtest itself.
Click to expand...

I lost faith in speedtests some time back and tend to rely on ping times and actual throughput nowadays. Pretty sure MyBB has been misreporting ping times to Cape Town for a while (PE routes thru Cape Town):

MyBB - http://speedtest.ookla.com/result/2738818531.png

Speedtest - http://www.speedtest.net/my-result/3853676636

Current stats with Radio4 iPlayer streaming & network under light load - 8mbps line, TI 10MB Uncapped (note download not running during speed & ping tests):

Ubuntu http download (FF + DownThemAll) Download speed.jpg

Code: 
[mick@mick1 ~]$ ping -qc 10 bbc.co.uk
PING bbc.co.uk (212.58.246.103) 56(84) bytes of data.

--- bbc.co.uk ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9174ms
rtt min/avg/max/mdev = 185.067/187.092/200.840/4.619 ms
 
Last edited:
Came across this:

http://www.cnet.com/how-to/how-to-find-a-reliable-network-speed-test/

Quote: "Many experts claim HTML5-based speed tests are more accurate than tests that use Java and Adobe Flash. Others point out that multithread tests such as those used by Ookla (Speedtest.net and branded by many ISPs) don't represent real-world network traffic as well as single-thread tests."

And this:

http://www.digitaltrends.com/computing/best-internet-speed-tests/

This is what is said about www.speedof.me and why I firmly believe that its results are more reflective of real internet conditions:

Quote: "and all files are downloaded and uploaded in sequence–rather than simultaneously–to imitate real internet browsing conditions."
 
^^Pretty much what i was trying to say just more eloquent :p
 
acp said:
This is what is said about www.speedof.me and why I firmly believe that its results are more reflective of real internet conditions:

Quote: "and all files are downloaded and uploaded in sequence–rather than simultaneously–to imitate real internet browsing conditions."
Click to expand...
Problem is speedof.me lacks a SA server so it's testing to London which might accurately reflect browsing The Telegraph or Guardian but not local news sites.
 
MickZA said:
Problem is speedof.me lacks a SA server so it's testing to London which might accurately reflect browsing The Telegraph or Guardian but not local news sites.
Click to expand...

Quite right, yet for me, its likely that >90% of my internet use is international, not local.
One can only stomach so many news24 articles per day... ;)

I specifically notice when streaming, YouTube and anything else. If speedof.me gives me good results - streaming happens in 720p.
Usually when streaming is borked, speedtest.net still tells me my line is flying, but speedof.me tells a different story...
 
acp said:
Usually when streaming is borked, speedtest.net still tells me my line is flying, but speedof.me tells a different story...
Click to expand...
Ahh, but if you were to test to London using speedtest.net you'd probably find it reports the same as speedof.me :)

... plus if speedof.me had a SA server it would report all was good - does speedof.me have an option to change servers or does it default to the closest?
 
MickZA said:
Ahh, but if you were to test to London using speedtest.net you'd probably find it reports the same as speedof.me :)

... plus if speedof.me had a SA server it would report all was good - does speedof.me have an option to change servers or does it default to the closest?
Click to expand...

Probably good that you used 'probably' ;)
I might not be that stupid...
Case in point, see below:
Both tests run after the other, London servers.
Don't think one can set default for speedof.me, for SA it defaults to London, therefore I set speedtest.net to London too.

3854078718.png

speedof.me_14-10-23.jpg
 
acp said:
Quite right, yet for me, its likely that >90% of my internet use is international, not local.
Click to expand...

Quite unlikely.

One can only stomach so many news24 articles per day... ;)

I specifically notice when streaming, YouTube and anything else. If speedof.me gives me good results - streaming happens in 720p.
Click to expand...

More than 98% of YouTube is local. YouTube doesn't feature on international traffic at all (p2p makes up the biggest portion of intl) where islt has the biggest slice of local traffic.
 
acp said:
Quote: "Many experts claim HTML5-based speed tests are more accurate than tests that use Java and Adobe Flash.
Click to expand...


Maybe only due to the fact that on a machine with a slower cpu flash-based tools will under report because it is limited by cpu. E.g. with a core i7, I can speedtest at 850Mbps,, with a core 2 duo only about 300Mbps.

Anyway, you can only compare accuracy between tools that are intended to measure the same thing. Speedtest is specifically designed to try and test maximum throughput (using multiple tcp connections). But, web browsing would be similar in some ways, with multiple connections to load a single site, while a non-p2p download would be different.

http://www.digitaltrends.com/computing/best-internet-speed-tests/

This is what is said about www.speedof.me and why I firmly believe that its results are more reflective of real internet conditions:

Quote: "and all files are downloaded and uploaded in sequence–rather than simultaneously–to imitate real internet browsing conditions."
Click to expand...

There is no easy way to test real-world speeds, because real-world internet use is very complex. Yes, you can have set measurements that measure e.g. real-world single-tcp connection throughput from a specific source, and compare them (we do that using some automated downloads from a few different sources to see the trends).

But, note that latency has quite a big impact on real-world throughput for single-connection traffic, international will always seem slower unless you have > 10 streams. That is why p2p and CDNs are both effective at providing better throughput, but p2p is abusive in that it floods buffers and connection tables and impacts less aggressive traffic severely.
 
bwana said:
I spoke to telkom today regarding my line speed and they said that even though I'm on a 10mbps line the ISP portion of my account (home uncapped prem) hasn't been upgraded so I'm still on 4mbps speeds. Fair enough but it got me wondering why, when I run a speedtest, am I consistently getting 8+mbps speeds?

View attachment 160623
Click to expand...

We have been investigating this, but can't yet explain it. But you probably won't see the same thing if you speedtest to SAIX.

Are they perhaps over-prioritising speedtests?
Click to expand...

No, it's not that ... there may be some some weird routing issue, as it affects our local speedtest servers and maybe mybroadband (in Gauteng), but not others.
 
Thanks ranger. What would a day be without learning something new - just because it says .com on the wrapper doesn't mean it is international right?
Could you help explain the difference between the different speed tests?

P2P - almost never use it, not worried about it, scheduled for 00h00-07h00 regardless.
The news servers I use are international and do seem to run at line speed most nights, then again, that is also backed-up by any speed test during those hours.

Just ran this test, and honestly, this is the browsing experience I'm having:
speedof.me_14-10-23 (1).jpg

Sadly, not this:

3854337259.png
 
Xzib1t said:
More concerning is my Nvidia driver download, which is peaked at 140 kbps.

Now it's averaging around the 50 kbps mark
View attachment 160809
Click to expand...

We are awaiting more transmission (in preparation for IPC upgrades for Cape Town) we initiated about 3 months ago when we ordered the upgrades that used the last capacity. Apparently more transmission cards had to be ordered from overseas, which have now arrived and been installed. Hopefully they should be completed next week. IPC upgrade may take a few more days.

The Cape Town site is currently worst off, Gauteng isn't great (also waiting for more transmission), Durban is fine (enough transmission for at least the next 4 months).
 
ranger said:
We are awaiting more transmission (in preparation for IPC upgrades for Cape Town) we initiated about 3 months ago when we ordered the upgrades that used the last capacity. Apparently more transmission cards had to be ordered from overseas, which have now arrived and been installed. Hopefully they should be completed next week. IPC upgrade may take a few more days.

The Cape Town site is currently worst off, Gauteng isn't great (also waiting for more transmission), Durban is fine (enough transmission for at least the next 4 months).
Click to expand...
Thanks man, I'll hold you to that.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X