Afrihost CAPPED ADSL Feedback (MTN)

Status
Not open for further replies.
I've made a few changes from my side, can you please restart your ADSL modem and check if there is any difference.

Great thank you for your help. I am not at home any more but will do so when I get back,...
 
Thanks Dude! :)

Including the region does actually help us in picking up any common issues a lot faster.

All of the packet loss issues have been in the South and the South IPC certainly doesn't have the best reputation, so would it be possible that it's an issue with the IPC during peak times when reaching full capacity?

In CS last night I was getting up to 85% loss. Completely unusable to say the least..My Web Africa account is almost capped so I'll be completely unable to play if it isn't fixed soon.

Also a reminder that pingplotter and pingtest don't show any packet loss (pingplotter to CS and TS servers). Have you guys got a copy of BF4 or CS to test it out late evening? That is how you will replicate the issues. Also very easy to see the loss on TS which gets worse and worse the longer you stay connected and try speak to people. If possible, your guys should test at around 8pm (in the SOUTH) as this is when it's at its worst.

Don't worry about what servers I experience this on, because it's all of them.
 
Western Cape,Mossel Bay/George area.
Packet loss problems at random during the day and at night it shoots to 20-30%
Capture.PNG

This is as we speak. when it's at this level i don't even bother gaming as its all stutter :(

Tested
-WA,No problems with packet loss or ping
-Crystal Web,No problems with packet loss or ping
-Mweb,No problems with packet loss or ping
 
Last edited:
Western Cape - Northern Suburbs - Bothasig/ Edgemead
High Packet Loss in the evenings and Slow torrent downloads. 10 MB line
I posted tracert data and speedtest results which showed nothing sinister.

My torrents only picked up speed after 2AM
 
All of the packet loss issues have been in the South and the South IPC certainly doesn't have the best reputation, so would it be possible that it's an issue with the IPC during peak times when reaching full capacity?

In CS last night I was getting up to 85% loss. Completely unusable to say the least..My Web Africa account is almost capped so I'll be completely unable to play if it isn't fixed soon.

Also a reminder that pingplotter and pingtest don't show any packet loss (pingplotter to CS and TS servers). Have you guys got a copy of BF4 or CS to test it out late evening? That is how you will replicate the issues. Also very easy to see the loss on TS which gets worse and worse the longer you stay connected and try speak to people. If possible, your guys should test at around 8pm (in the SOUTH) as this is when it's at its worst.

Don't worry about what servers I experience this on, because it's all of them.

I think it's really important to test a specific scenario so the more we can replicate a specific set of circumstances.
 
Just in general, it seems like most of the complaints of speeds and packet loss are from the South so I am going to log a general query to find out if there is anything that I can report.

I promise I'll post any feedback I get if there is a general problem that we find.
 
CSGO new day.jpg

Tested it quick. 10:49am, Cape Town, Southern Suburbs.

I didn't realize packet loss had gotten so bad during off peak :/

Can confirm that pinging the exact server gives no packet loss. Same with mumble.
 
We are seeing pretty high demand in the South at the moment. This seems to have spiked uncharacteristically from previous trends, and we're investigating to see what might have increased traffic so dramatically. I don't know of any big software releases at the moment (I know there was a very small iOS 8.2 update released in the last day or so).

We're still busy investigating, but this also appears to be affecting traffic in the East, so we need to look into exactly where the additional demand originates.

I'll keep you posted on what we hear back from the team, but so far it seems like it's demand that we need to manage to normalise traffic on the network.
 
Awesome! So I decide to downgrade my line right. I go through the process, and afterwards I get an email stating that there will not be a credit applied for downgrading. Fine I decide to then do it closer to the billing run, and ask Afrihost to cancel the downgrade. The person tells me I can't as the process is already with Telkom!!

So remind me again why I pay in advance if you cannot apply a credit note when I downgrade?
 
Still getting 5% loss on TS at the moment. Nothing has changed this side

We were seeing high demand about an hour ago, this seems to have normalised in the last 20-30 minutes. Latency, throughput and packet loss should be normal in general.

Are you getting packet loss anywhere else?
 
Awesome! So I decide to downgrade my line right. I go through the process, and afterwards I get an email stating that there will not be a credit applied for downgrading. Fine I decide to then do it closer to the billing run, and ask Afrihost to cancel the downgrade. The person tells me I can't as the process is already with Telkom!!

So remind me again why I pay in advance if you cannot apply a credit note when I downgrade?

we don't generally give credit for line downgrades, but drop me a PM and let me see what I can do for you :(

Oops, just realised I have disabled PM's cos I've leaving for the US in a few hours. Mail me on [email protected] and I'll make sure we take proper look :)
 
We were seeing high demand about an hour ago, this seems to have normalised in the last 20-30 minutes. Latency, throughput and packet loss should be normal in general.

Are you getting packet loss anywhere else?

Haven't checked CS, but generally when I see loss on TS, it's worse in CS.

I don't know a whole lot about networks, but seeing as you say there has been an increase in demand, that basically explains it. It is at its worst during peak times and goes away after midnight. Your management software should be protecting network integrity at the cost of throughput, not packet loss. Or am I wrong?
 
Haven't checked CS, but generally when I see loss on TS, it's worse in CS.

I don't know a whole lot about networks, but seeing as you say there has been an increase in demand, that basically explains it. It is at its worst during peak times and goes away after midnight. Your management software should be protecting network integrity at the cost of throughput, not packet loss. Or am I wrong?

Yes, you're right, the network management preserves latency over throughput, for the best overall experience. However, it is reactive to demand and then there is a short period when the system reviews whether changes made are effective enough (to prevent a "clamp down" which may give client's a poorer experience given the resources available).

From what I've seen, this was an unexpected spike in demand today, so it took us a while to bring it under control, but seems to be stable now.

IN terms of packet loss, that should not be happening. If it is, we need to know exactly where on the network packets are dropped, so we can address whether this is a hardware failure, routing or network management problem or something else.
 
Yes, you're right, the network management preserves latency over throughput, for the best overall experience. However, it is reactive to demand and then there is a short period when the system reviews whether changes made are effective enough (to prevent a "clamp down" which may give client's a poorer experience given the resources available).

From what I've seen, this was an unexpected spike in demand today, so it took us a while to bring it under control, but seems to be stable now.

IN terms of packet loss, that should not be happening. If it is, we need to know exactly where on the network packets are dropped, so we can address whether this is a hardware failure, routing or network management problem or something else.

Is there anything we can do to help track down the problem? Or do your guys know roughly where to look now?
 
Is there anything we can do to help track down the problem? Or do your guys know roughly where to look now?

We've managed to isolate why demand appeared to peak and that seems to be under control now. But in terms of packet loss, we need to see test results showing exactly where in the routing packet loss occurs. In game reports don't tell us enough. For us to take action, it needs to be on a router or server within our network.
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X