Telkom Internet Capped/Uncapped User Feedback (Pt2)

CataclysmZA said:
I'm diagnosing a speed/latency issue for a neighbour who just signed up with Telkom Internet. Are others experiencing this as well?

Speed isn't an issue, because YouTube tests ran just fine, but actual speed tests, browsing, and some HTTP downloads are really, really bad.

View attachment 448536
Click to expand...
Please test with pingplotter or winmtr.

The problem with traceroute, as exhibited here, is that it test progressively along the path, and it's results are only actually useful when there is no variability in utilisation of any link in the path for the full duration of the traceroute.

At least mtr and winmtr do the following (as far as I know)
- first determine the path
- at regular intervals ping all hops on the path at exactly the same time
- provide the min/avg/max/variance for each hop

This makes it much easier to see what the source if the latency was.

(However, from just the traceroute I would say there is a >90% probability that *all* of the latency is introduced at the DSLAM, because it should be less than 30ms, if it was 200ms it is starting to delay packets and it only takes a slight bit more traffic to start dropping packets. In this case, changing ISPs won't donanything and the ISP can't directly do anything, the entity that sells the DSL line to the customer at present could try and get the DSLAM backhaul upgraded by Openserve, but as far as I know all DSLAMs that are congested have been upgraded as far as the backhaul technology - ATM - allows, so migration to ME backhaul and/or MSAN and/or FTTH will be the only solution)
 
I would have tested with either, but the line was slow enough to make getting Pingplotter a chore. I'll be going back there later to see. Armed with those apps.

The exchange my neighbour is connected to is the same as mine, located in C-Place, Jeffreys Bay (EJFB1 or something like that), and my line is perfectly fine. I'll report back later.
 
CataclysmZA said:
I would have tested with either, but the line was slow enough to make getting Pingplotter a chore. I'll be going back there later to see. Armed with those apps.

The exchange my neighbour is connected to is the same as mine, located in C-Place, Jeffreys Bay (EJFB1 or something like that), and my line is perfectly fine. I'll report back later.
Click to expand...
Exchange!=DSLAM.

Some DSLAMs in an exchange may be congested while others are not. In this case it may be possible to have the line moved to another DSLAM if there are ports available on a non-congested DSLAM. But ... first you would have to get past the "contact centre" ...
 
ranger said:
Exchange!=DSLAM.

Some DSLAMs in an exchange may be congested while others are not. In this case it may be possible to have the line moved to another DSLAM if there are ports available on a non-congested DSLAM. But ... first you would have to get past the "contact centre" ...
Click to expand...

hehehehe - good luck!!

rather phone a friend! :D
 
ranger said:
Exchange!=DSLAM.

Some DSLAMs in an exchange may be congested while others are not. In this case it may be possible to have the line moved to another DSLAM if there are ports available on a non-congested DSLAM. But ... first you would have to get past the "contact centre" ...
Click to expand...

Right, right. Not the same port, not even on the same rack, probably.

I do see the local techies driving around, so if things don't improve, I'll get their issue logged and see if the local guys respond.
 
Morning. I am having massive issues with my ADSL line. I am unable to Skype, steam anything other than youtube or use teamviewer.

Please would someone be able to help me interpret the results of this pingplotter diagram. It seems like for about 30 seconds at a time my home ADSL live experiences 100% packet loss in cycles. I have run the same test on the propery right next door which also has an ADSL line and there are no issues at all. Both sites are almost exactly the same distance form the exchange. The only difference is the line with an issue is 4MB and the second business line is a 2MB line. I have not had issues in the past with being on a 4MB line.

TelkomPVT.jpg


Thanks very much
 
I am on Telkom 40GB /4MB capped service. I notice that for the last two months (possibly even more) that downloads after midnight goes off my cap. What gives? Don't they have the night surfer anymore?
 
isie said:
Anecdotal or did you get this from the customer portal?
Click to expand...

A combination. My normal surfing hardly amounts to much data usage. And all downloads are starting up at 00h10 and ends way before 07h00. But yet my cap is almost done for the month. And last month I exceeded my cap. Something is not right.

I logged onto this new Telkom portal just now. Under Open Connections it shows my log on time as 2h20 four days ago!! So in the last four days my line never went down? And I would guess all activity would have been counted as daily? Or am I confused?

telkUntitled.png
 
SeRpEnT said:
A combination. My normal surfing hardly amounts to much data usage. And all downloads are starting up at 00h10 and ends way before 07h00. But yet my cap is almost done for the month. And last month I exceeded my cap. Something is not right.

I logged onto this new Telkom portal just now. Under Open Connections it shows my log on time as 2h20 four days ago!! So in the last four days my line never went down? And I would guess all activity would have been counted as daily? Or am I confused?

View attachment 449208
Click to expand...

your line never went down would not affect your usage, above this should have your usage history graph check if any night data usage there.
 
isie said:
your line never went down would not affect your usage, above this should have your usage history graph check if any night data usage there.
Click to expand...

Yes I can see the graph with "fixed" vs night surfer downloads per day. Which shows that night surfer usage has been much lower than fixed usage for many days. While in reality its pretty much impossible as normal daily internet usage is just surfing 99% of the time. All D/L are during night surfer hours. Without proper daily hour info this graph is quite meaningless imo.

But let's say there was some kind of background data usage that I am not aware of. Which culprits should I investigate? We have two cellphones on the wifi and one tablet. And my wife connects the laptop for work (only emails). Windows 10 updates recently? Rogue android apps using background data? I know I can track android data usage, but is there a way to monitor Windows data usage without installing a 3rd party program? Thanks in advance.
 
SeRpEnT said:
Yes I can see the graph with "fixed" vs night surfer downloads per day. Which shows that night surfer usage has been much lower than fixed usage for many days. While in reality its pretty much impossible as normal daily internet usage is just surfing 99% of the time. All D/L are during night surfer hours. Without proper daily hour info this graph is quite meaningless imo.

But let's say there was some kind of background data usage that I am not aware of. Which culprits should I investigate? We have two cellphones on the wifi and one tablet. And my wife connects the laptop for work (only emails). Windows 10 updates recently? Rogue android apps using background data? I know I can track android data usage, but is there a way to monitor Windows data usage without installing a 3rd party program? Thanks in advance.
Click to expand...

If you use Windows, have you checked update history to see if any updates have been installed?
 
Pingplotter tests conducted last night. Same device (my Windows 10 notebook), with all other devices turned off, no contention for bandwidth on the local network.

LMUiyWx.jpg

evsVoO7.jpg

uHovVDi.jpg


Local browsing is OK, but mostly slow. Facebook's UI hangs quite a lot due to the latency. YouTube tests run just fine, goes 90% full speed most of the time, looking at the video stats. Changing to Google's DNS resolves some issues, but not all of them.
 
Last edited:
CataclysmZA said:
Pingplotter tests conducted last night. Same device (my Windows 10 notebook), with all other devices turned off, no contention for bandwidth on the local network.

LMUiyWx.jpg

evsVoO7.jpg

uHovVDi.jpg


Local browsing is OK, but mostly slow.
Click to expand...

You need to try some content that isn't on MWEB's network (maybe from http://mirror.ac.za or http://ftp.sun.ac.za , or some other ISP, maybe sites hosted at Afrihost).

From the screenshots, the only thing that stands out is that the latency to news24 has a 50ms increase from where it is handed over to MWEB at Teraco CT1 until it gets to news24 over MWEB's network, indicating that QoS policies are kicking in. This would be MWEB's issue, they should be doing adequate capacity planning for their traffic so they don't shape traffic on the links to their peering router(s).

Facebook's UI hangs quite a lot due to the latency.
Click to expand...

Facebook has deployed some local CDN nodes that Telkom Internet has access to, so media assets shouldn't be too slow, but the timeline itself (and other dynamic contenr) is only served from bigger sites (all overseas), so it will be slower. The 250ms isn't too bad for international sites (considering the ~ 45ms to the IPC link).

YouTube tests run just fine, goes 90% full speed most of the time, looking at the video stats. Changing to Google's DNS resolves some issues, but not all of them.
Click to expand...

Using Google DNS could only possibly help cases where some sites don't resolve, if you experience any persistent issues, please report them. But using Google DNS could result in you not using CDNs Google isn't aware of. TI had access to a few others, some that other local ISPs may not have.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X