Afrihost Business Uncapped Feedback

[TEXTILE GUY]

We received feedback from MTN Business, & the intermittent drops in speeds are linked to capacity being reached during certain hours on our Eastern IPC.
We hope to have this resolved soonest.

A temporary solution will probably be shaping for uncapped users, a long term solution will be to increase capacity on the IPC.

Just to be clear on business accounts and shaping .... This from AH technical responding to my BUSINESS account.^^

BUSINESS ACCOUNTS ARE SHAPED THEN

If they were not I would not be running at 3 megs on a 4 meg line. And if '' uncapped accounts'' are being shaped to resolve the problem how come my problem is not resolved???

AH are really talking in circles lately ....

 

[requiem]

Just to be clear on business accounts and shaping .... This from AH technical responding to my BUSINESS account.

BUSINESS ACCOUNTS ARE SHAPED THEN

This has been demonstrated many times. Each time though the official word is that we are confused, don't understand basic technical concepts, and it must be a problem on our side.

 

[AfriMan]

View attachment 129425


Zertop's "Is it my Line" Results
Basic Report:

Your average ping was 16ms
Your maximum ping was 82ms

Looking at your average ping, your line seems to be running perfectly
Looking at your maximum ping, however, it seems as though there could be a serious intermittent fault on the line. This could be a cause of an issue. Please post the results (at the end of the program) into the forum for advice.


Detailed Report

Ping results for last hop before account switches to ISP's hardware (Hop 2 on Tracert) 

Pinging 105.236.11.193 with 32 bytes of data:
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=14ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=12ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254
Reply from 105.236.11.193: bytes=32 time=14ms TTL=254
Reply from 105.236.11.193: bytes=32 time=42ms TTL=254
Reply from 105.236.11.193: bytes=32 time=82ms TTL=254
Reply from 105.236.11.193: bytes=32 time=13ms TTL=254

Ping statistics for 105.236.11.193:
    Packets: Sent = 30, Received = 30, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 12ms, Maximum = 82ms, Average = 16ms
 
2014/07/02 18:02:37.42

I still cannot stream a Youtube vide at 720p without it buffering to hell.

I think that spike in latency might show that there is intermittent latency, which might explain buffering. Unless you don't experience these latency jumps on other networks.

 

[Bursty-dude]

I don't have any other test accounts on any other networks.

 

[AfriMan]

What a total load of bull****.

So if 1 user is having an acceptable experience - it is not the network? You are completely loopy.

Everytime you have been provided proof of packet loss within MTN's network, you deny deny deny.... and wait for the problem to go away.


I am disgusted at how blatantly dishonest Afrihost has become.

Sorry to hear that you're disgusted, really not my intention

I'm merely saying that if 1 client can route without latency/buffering, it means it is then possible and not a design limitation on the network. So simply drawing conclusion that the network is broken would spend a lot of fruitless time looking in the wrong place.

It would be more pertinent to look at those reporting the same symptoms and look for commonalities. Are they on the same exchange? Are they using the same router model? Are they using the same DNS settings? This will most likely be a much more constructive process.

We definitely need your help to find out what is happening, so we need you to test so we can quantify what you are experiencing and match them to the circumstances you are connecting under.

 

[AfriMan]

My line was perfect until just over 2 weeks ago.
I could have a download running in the background AND watch a Youtube vid at 720 or 1080p at the same time.
Now I have to stop all downloads and wait for the vid to buffer for at least 10 minutes before watching.

But not to worry, there are no faults anywhere in the Matrix. It is all just our imagination causing ripples in the space-time continuum known as AH ADSL Post MTN......

Sounds like we definitely need to test further here. Drop me a PM if you can so we can run comparative tests to recreate the problem.

 

[AfriMan]

Just to be clear on business accounts and shaping .... This from AH technical responding to my BUSINESS account.^^

BUSINESS ACCOUNTS ARE SHAPED THEN

If they were not I would not be running at 3 megs on a 4 meg line. And if '' uncapped accounts'' are being shaped to resolve the problem how come my problem is not resolved???

AH are really talking in circles lately ....

I think the issue is content here. I don't think the person responding is referring to Uncapped accounts (not Business and Capped) which are shaped to improve overall throughput and realtime performance on the network.

We are now shaping Uncapped accounts from 7:30am to 10pm to ensure Business and Capped performance. So if your throughput is still poor (as we are definitely seeing improvement in our tests) then we'll need to do some proper troubleshooting on this.

 

[AfriMan]

This has been demonstrated many times. Each time though the official word is that we are confused, don't understand basic technical concepts, and it must be a problem on our side.

We've been very open that when we've been in need of additional capacity that there has been contention. But we've never shaped Business and we don't plan on doing this.

 

[requiem]

Sorry to hear that you're disgusted, really not my intention

I'm merely saying that if 1 client can route without latency/buffering, it means it is then possible and not a design limitation on the network. So simply drawing conclusion that the network is broken would spend a lot of fruitless time looking in the wrong place.

It would be more pertinent to look at those reporting the same symptoms and look for commonalities. Are they on the same exchange? Are they using the same router model? Are they using the same DNS settings? This will most likely be a much more constructive process.

We definitely need your help to find out what is happening, so we need you to test so we can quantify what you are experiencing and match them to the circumstances you are connecting under.

If 100 users are trying to use your network. 99 of them report lag, spikes, problems etc... but 1 screams how happy he is.

What is the common problem?

Each time we have shown the "intermittent" latency on the 3rd / 4th MTN hop - the common response is "I don't see a problem from my side. It might be an exchange issue / insert other reason..."


It is absolutely pointless, to log problems anymore, as when you do, the ticket is logged. Your connection is closed, and you are told to try again. One spends 15 - 20 minutes, gathering all the information. Finally the support desk just ignores the problem until you go away when you get too tired to keep on fighting.

So. Here I challenge you to give me a straight and useful answer.

While using Afrihost -

While using WebAfrica.

Do you see the jumps to 63 / 73 ms on the local loop?

This normally happens with packet loss, is behaving itself so far this morning in that regard though.

Now please tell me why the problem is on my line?

 

[AfriMan]

If 100 users are trying to use your network. 99 of them report lag, spikes, problems etc... but 1 screams how happy he is.

What is the common problem?

Each time we have shown the "intermittent" latency on the 3rd / 4th MTN hop - the common response is "I don't see a problem from my side. It might be an exchange issue / insert other reason..."


It is absolutely pointless, to log problems anymore, as when you do, the ticket is logged. Your connection is closed, and you are told to try again. One spends 15 - 20 minutes, gathering all the information. Finally the support desk just ignores the problem until you go away when you get too tired to keep on fighting.

So. Here I challenge you to give me a straight and useful answer.

While using Afrihost -

While using WebAfrica.

Do you see the jumps to 63 / 73 ms on the local loop?

This normally happens with packet loss, is behaving itself so far this morning in that regard though.

Now please tell me why the problem is on my line?

It looks to me like both results are intermittent. There are latency spikes across the board, so it doesn't really demonstrate that one network performs significantly better than the other. Also, if your line is experiencing sync loss, it may perform progressively worse, but then well again for a short period when you switch accounts and re-sync to the exchange.

There are a range of factors that could be coming into this, so I don't think we've really defined the problem fully. If you are willing, we can test with you and run some thorough troubleshooting on this. If you are willing to do that, I'd be very grateful to you.

Drop me a PM with your details so we can get started.

 

[requiem]

It looks to me like both results are intermittent. There are latency spikes across the board, so it doesn't really demonstrate that one network performs significantly better than the other. Also, if your line is experiencing sync loss, it may perform progressively worse, but then well again for a short period when you switch accounts and re-sync to the exchange.

There are a range of factors that could be coming into this, so I don't think we've really defined the problem fully. If you are willing, we can test with you and run some thorough troubleshooting on this. If you are willing to do that, I'd be very grateful to you.

Drop me a PM with your details so we can get started.

The latency spike on Web Africa was on 1 hop from 8 (avg) to 28 (max)

The latency spikes on Afrihost was on 3 hops

from 8 (avg) to 63 (max)
from 9 (avg) to 62 (max)
from 10 (avg) to 73 (max)

Hardly fair to say "across the board".


I have also stated that streaming on Afrihost is a disaster. Gaming on Afrihost is a disaster. When using webafrica everything just works.

What possible other result interpretation could there be?

 

[AfriMan]

The latency spike on Web Africa was on 1 hop from 8 (avg) to 28 (max)

The latency spikes on Afrihost was on 3 hops

from 8 (avg) to 63 (max)
from 9 (avg) to 62 (max)
from 10 (avg) to 73 (max)

Hardly fair to say "across the board".


I have also stated that streaming on Afrihost is a disaster. Gaming on Afrihost is a disaster. When using webafrica everything just works.

What possible other result interpretation could there be?

Looking at the variance, it is fair to say that there is intermittent latency spiking. There could be many contributing factors and circumstances at play here, so I do think it's premature to make a diagnosis without further tests.

 

[requiem]

Forwarded the required tests to critical care early this afternoon.


As you can imagine, no one has gotten back to me or bothered to reply.

 

[TEXTILE GUY]

. So if your throughput is still poor (as we are definitely seeing improvement in our tests) then we'll need to do some proper troubleshooting on this.

Sorry, not to be funny or be otherwise .... does this imply that the past round of tests with SAIX ( some 7 tests on my side) and e-mails were not ''proper''?
You did say you had escalated this to MTN last week ......

I dont want to drag AH thru the mud - I remain loyal, please just give me a consistent service.
( And PLEASE dont give me reason to phone in - your call centre is lacking at best - use this or dont).

I am really asking nicely.

 

[AfriMan]

Forwarded the required tests to critical care early this afternoon.


As you can imagine, no one has gotten back to me or bothered to reply.

It will obviously take some time to analyse your results and try to recreate what they can on our side to establish where they think the problem could be. Let's give them some time to get back to you

 

[AfriMan]

Sorry, not to be funny or be otherwise .... does this imply that the past round of tests with SAIX ( some 7 tests on my side) and e-mails were not ''proper''?
You did say you had escalated this to MTN last week ......

I dont want to drag AH thru the mud - I remain loyal, please just give me a consistent service.
( And PLEASE dont give me reason to phone in - your call centre is lacking at best - use this or dont).

I am really asking nicely.

I dunno if you've been testing with AfriGuy, my doppelganger. But drop me a PM or an email (afriman@afrihost.com) and we'll go through all the tests you've done together and see what we can see.

 

[requiem]

I dunno if you've been testing with AfriGuy, my doppelganger. But drop me a PM or an email (afriman@afrihost.com) and we'll go through all the tests you've done together and see what we can see.

So the response from CRITICALCARE is

"From my end, using the Afrihost capped account, this morning I am not seeing any packet loss, maybe when the network is busy.
Then that could be contention. "

I.e. we are not going to help you whatsoever, nor listen to you when you describe the problem. We will simply deny there is anything wrong.

 

[AfriMan]

So the response from CRITICALCARE is

"From my end, using the Afrihost capped account, this morning I am not seeing any packet loss, maybe when the network is busy.
Then that could be contention. "

I.e. we are not going to help you whatsoever, nor listen to you when you describe the problem. We will simply deny there is anything wrong.

I think it is fair to say that we must be able to recreate the problem to say there is a problem on the network. If the network is at fault, we would see packet loss using your account on any DSL line (or a similar account using the same network rules). It again comes down to systematic troubleshooting and not making assumptions too early. I was testing streaming on Capped and Uncapped (both shaped and unshaped) and I have no packet loss or buffering in general (up to 720p on a 4Mbps synched line).

 

[TEXTILE GUY]

I dunno if you've been testing with AfriGuy, my doppelganger. But drop me a PM or an email (afriman@afrihost.com) and we'll go through all the tests you've done together and see what we can see.

It was with your critical care team, some guy called Oscar if that helps.

 

[AfriMan]

It was with your critical care team, some guy called Oscar if that helps.

If you're with Critical Care, you are really in the best hands within the company to get a solution. They are a special unit that is setup specifically to get clients through difficult (and complex) issues like this. Please give Oscar a chance, I promise he will come good