Afrihost Uncapped ADSL Feedback

Status
Not open for further replies.
I wonder how much all these issues can be attributed to Telkoms account upgrades and their inability to accommodate the extra bandwidth requirements?? Ultimately, at the end of the day the buck always stops at their front door and we know how efficient and effective they are!!
This type of throttling by the ISP means that it is at the ISP's door...
 
My usage last night:

View attachment 27114

From 1am to 2am it was 1Mbit.
From 2am to 3am it slowly creeped up to 4Mbit
From 3am to 4:30am it remained at roughly 4Mbit
From 4:30am to 6:30am it jumped to and remained at 10Mbit (maintaining above 1000kb/s which is more than I'm used to).
Currently it's at 4Mbit again.

This shaping or throttling nonsense is the most irritating inconsistent mess ever.

View attachment 27116

View attachment 27118

2289350374.png


So inconsistent!!!

Johannesburg... London... Johannesburg. Seems pretty consistent to me for those locations.
 
I have received an automated response to a mail I sent them earlier, asking for an explanation regaring their Fair Usage Policy, whether we are being shaped and/or throttled, and when we can expect a resolution to the current problems.
No reply as yet other than the automated response saying that someone will contact me shortly.

Plugg seems more attractive with every hour that passes.
 
"so that browsing, FTP and VoIP will get best throughput over downloads."

Who uses FTP now days? VoIP - Good, although latency is the most important aspect here, speaking of which I see it is all over the place this morning so obviously still tweaking going on. BW is back to normal though.
 
Guys, this is not shaping, its THROTTLING. If I didn't know better, I'd say Afrihost's technical department don't actually KNOW how to setup & manage shaping models on even a basic level.

I've been singing Afrihost's praises for a while now, even during the transition to MTN (I was one of the lucky low-ish latency customers), but since this last weekend things have just gotten silly.

Mweb's 6-month contract special (with 2 months free) on their 10mb uncapped bundle works out to roughly the same as Afrihost's 10mb uncapped bundle (R1359*6 - R2718 = R906pm vs. R897pm). Even though this price is over the first 6 months, I'm still VERY tempted to move back.

AFRIHOST, STOP THROTTLING THE INNOCENT USERS AND GET YOURSELVES INTO SHAPE

(see what i did there)
 
Afrihost made a lot of friends with the excellent performance and value of these accounts from May until the switch to the MTN network. I think many of us still want to believe that the performance we got on the IS network is the way Afrihost wants these accounts to run and are prepared to wait a few days to see if they get restored to what they were before the switch. Unfortunately we have all paid for the month so cancelling now will only take effect from next month. Surely Afrihost must realize that they risk a flood of cancellations closer to the end of the month if they don't do something urgent. I am all for punishing a service provider for bad service and fully intend to cancel before the end of the month if it stays like this. Does anyone know if it is possible to cancel now and then perhaps if normal service is restored to stop the cancellation via client zone?

You got to decide when you are going to stop supporting these kind of practices, ask yourself... when will this behavior change if people just keep on being wishful?***

You are saying that it is ok not to do pretesting, but in fact preferable to use yourself as a test case.

If this has been done once tell me what is stopping them from doing it again especially if they bargain and know that people like yourself will hold out. Surely money isn't something you can just toss away? I would at least ask a refund for each month it has been crap. ANY decent company would offer compensation for their ****ups.


Simple question: What exactly stopped them from pretesting the network themselves AND bringing over a few to test?


***Case and point... tell me how great WA's uncapped is since last year Nov 2011?
 
Last edited:
YSimple question: What exactly stopped them from pretesting the network themselves AND bringing over a few to test?

To be fair they did do this with the Rage vouchers and a few MyBB users (me being one of them). Of course there was no way for us to know QoS hadn't been setup when we were testing.
 
Simple question: What exactly stopped them from pretesting the network themselves AND bringing over a few to test?

They probably did do just that. The problem is we told them their network would fall to crap when they moved to MTN, they said just watch this space.
 
Guys, this is not shaping, its THROTTLING. If I didn't know better, I'd say Afrihost's technical department don't actually KNOW how to setup & manage shaping models on even a basic level.

I've been singing Afrihost's praises for a while now, even during the transition to MTN (I was one of the lucky low-ish latency customers), but since this last weekend things have just gotten silly.

Mweb's 6-month contract special (with 2 months free) on their 10mb uncapped bundle works out to roughly the same as Afrihost's 10mb uncapped bundle (R1359*6 - R2718 = R906pm vs. R897pm). Even though this price is over the first 6 months, I'm still VERY tempted to move back.

AFRIHOST, STOP THROTTLING THE INNOCENT USERS AND GET YOURSELVES INTO SHAPE

(see what i did there)

I see what you did there.
 
I see ping times are completely useless again. An average of 256ms to my uplink, which is normally 11-12ms, so 240+ms extra latency. Totally worth the money I pay to Afrihost every month.

$ ping -c 10 105.236.3.129
PING 105.236.3.129 (105.236.3.129) 56(84) bytes of data.
64 bytes from 105.236.3.129: icmp_req=1 ttl=253 time=288 ms
64 bytes from 105.236.3.129: icmp_req=2 ttl=253 time=257 ms
64 bytes from 105.236.3.129: icmp_req=3 ttl=253 time=256 ms
64 bytes from 105.236.3.129: icmp_req=4 ttl=253 time=240 ms
64 bytes from 105.236.3.129: icmp_req=5 ttl=253 time=249 ms
64 bytes from 105.236.3.129: icmp_req=6 ttl=253 time=251 ms
64 bytes from 105.236.3.129: icmp_req=7 ttl=253 time=253 ms
64 bytes from 105.236.3.129: icmp_req=8 ttl=253 time=255 ms
64 bytes from 105.236.3.129: icmp_req=9 ttl=253 time=258 ms
64 bytes from 105.236.3.129: icmp_req=10 ttl=253 time=262 ms

--- 105.236.3.129 ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9010ms
rtt min/avg/max/mdev = 240.692/257.519/288.516/11.753 ms
 
Did a speedtest now from the myBB speedtest to international

Download Speed: 43 kbps (5.4 KB/sec transfer rate)
Upload Speed: 354 kbps (44.3 KB/sec transfer rate)
Latency: 400 ms

Afrihost 2Mb uncapped.

And i was specifically looking for something that works during the day, not a 6pm to 7am account.
 
They probably did do just that. The problem is we told them their network would fall to crap when they moved to MTN, they said just watch this space.

To be fair they did do this with the Rage vouchers and a few MyBB users (me being one of them). Of course there was no way for us to know QoS hadn't been setup when we were testing.

Lets assume they did, clearly they didn't using CT users and QoS ?
 
We know things have been up and down this morning. So without going into the merits, we know about the problem and we are working on this.

Let me rather give you an idea of what is being done. We have MTN's network people working in our offices now, as well as some guys from a 3rd party company who are helping QA the shaping and QoS - they have experience in implementing this for other ISP's. We have a think tank set up in our boardroom, and this is being overseen by our directors, who are demanding a better result. We are also getting input from our Dev team, and our General manager who manages the support channels and has a good grasp of what the current complaints relate to. We have also engaged an external company, based in JHB and CT, who are now running intensive diagnostics on the network. they are running capped and uncapped accounts alongside our major competitors and analyzing the differences in performance, including latencies and QoS, and this is being fed back into the team.

As I've said before, no expense is being spared to address this problem, and to get things to where they should be. We have the full backing of MTN to make the changes we need to the network, and policies are being implemented and adjusted in real time.

Guys, Gian and the other directors are not going to stick us with a lemon network! They have confidence that this will be resolved, and as I said the message is clear from them that this needs to be resolved as top priority - no matter what the cost! Having said that, it's an ongoing process. I expect there will be ups and downs over the next few days of trial and error, but remember we were on a completely unshaped network since the 16th, as we didn't want to go straight into this process.
 
This type of throttling by the ISP means that it is at the ISP's door...

Point is do we really know what it is?

My 2 10mb uncapped accounts with Cybersmart, one in Kempton, the other in Norwood. Same router, W8960N both with attainable rates well over 10mb...




 
As I've said before, no expense is being spared to address this problem, and to get things to where they should be.

If no expense is being spared, you could, you know, add enough capacity to your network to handle the amount of traffic you have. Novel concept, I know.
 
Let me rather give you an idea of what is being done. We have MTN's network people working in our offices now, as well as some guys from a 3rd party company who are helping QA the shaping and QoS - they have experience in implementing this for other ISP's. We have a think tank set up in our boardroom, and this is being overseen by our directors, who are demanding a better result. We are also getting input from our Dev team, and our General manager who manages the support channels and has a good grasp of what the current complaints relate to. We have also engaged an external company, based in JHB and CT, who are now running intensive diagnostics on the network. they are running capped and uncapped accounts alongside our major competitors and analyzing the differences in performance, including latencies and QoS, and this is being fed back into the team.

As I've said before, no expense is being spared to address this problem, and to get things to where they should be. We have the full backing of MTN to make the changes we need to the network, and policies are being implemented and adjusted in real time.

Wow, that sounds like a proper response to the issue. Hopefully it pays dividends ASAP.
 
Johannesburg... London... Johannesburg. Seems pretty consistent to me for those locations.

Not at all.

4.77 on the first test to Jhb, 5.3 on the second. When I was with WA, it was consistently on 7.7 and when I moved to AH IS uncapped, it was consistently 8.

London was never under 4/5 for me.
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X