Afrihost Uncapped ADSL Feedback (Pt2)

Status
Not open for further replies.
Sorry to hear about this incident - I'll check this ticket out to see what exactly happened here.

In your ticket, I only looked at the product and not the billing aspect. I can appreciate that our guys did not give you the best service, but they are always genuinely doing their best to do what's best for our clients.

Will respond to you by PM with more feedback on this :(

Thanks, much appreciated :)
 
when does this happen ? - i seriously need to plan my way forward from this service that as far as i am concerned just evaporated on the first of this month

Ditto, not all of us have R20000 for a December vacation, so downloading and heavy streaming while at home is a must.

Edit:

Afriman/Afriguy/AfriGal, how wil shaping be on a 2mbps account? On 4mbps I get 115KB/s at most on a HTTP download, can I expect 100KB/s at the least on a 2mbps std uncapped account with heavy shaping?
 
Ditto, not all of us have R20000 for a December vacation, so downloading and heavy streaming while at home is a must.

Edit:

Afriman/Afriguy/AfriGal, how wil shaping be on a 2mbps account? On 4mbps I get 115KB/s at most on a HTTP download, can I expect 100KB/s at the least on a 2mbps std uncapped account with heavy shaping?

It depends on the policy that you're on. While we're aiming to alleviate shaping to a large degree - there may still be shaping present at some stages - it's a necessary component of managing an Uncapped service. The percentage shaping levels are standard across all line speeds though.
 
I finally have some clarity on the problem I was having.

Telkom engineer gave me a buzz now and reported that the ESR was indeed congested. Now, this will NOT give a high ping/latency on your trace route! As in my case my line was rate limited when QoS kicked in at the dslam itself.

It has been referred to the UMC division at Telkom to make them aware of the congestion. They weren't aware of it otherwise in part due to the line rate limiting QoS employed which keeps latency low, which in turn is what they use to identify congestion.

Now I at least know when my speed get's quartered it is the dslam QoS. Will have to wait for them to upgrade my area to fix this.

I'd like to thank AfriMan for his part in chasing this up with Telkom and getting the ball rolling for me to get an answer.
 
I finally have some clarity on the problem I was having.

Telkom engineer gave me a buzz now and reported that the ESR was indeed congested. Now, this will NOT give a high ping/latency on your trace route! As in my case my line was rate limited when QoS kicked in at the dslam itself.

It has been referred to the UMC division at Telkom to make them aware of the congestion. They weren't aware of it otherwise in part due to the line rate limiting QoS employed which keeps latency low, which in turn is what they use to identify congestion.

Now I at least know when my speed get's quartered it is the dslam QoS. Will have to wait for them to upgrade my area to fix this.

I'd like to thank AfriMan for his part in chasing this up with Telkom and getting the ball rolling for me to get an answer.

I wish there was more that we could do to help here - but this particular practice is very concerning. I hope that Telkom can figure out a better way to relieve congestion becuase it would cause mass confusion :(
 
I wish there was more that we could do to help here - but this particular practice is very concerning. I hope that Telkom can figure out a better way to relieve congestion becuase it would cause mass confusion :(

As am led to believe this is not the norm and is isolated to dslams connected to ESR's and NOT ASR's. Now I don't understand the technical side of it well myself, I do know that they weren't ware of the congestion on my particular dslam due to this though. When they investigated manually they could pick it up.

And yes. Hopefully they put into place a monitoring system specifically for these type of dslams as I am sure that I am not the only case like this.

A way to ID this is to get a tool that can track your bandwidth through put such as Bitmeter OS. I could clearly see that my speeds were indeed being CAPPED from 4 to 1 with pings staying low. Again - this is NOT an Afrihost problem. Just something else for many other people with problems to consider...congestion comes in many forms it seems.
 
It depends on the policy that you're on. While we're aiming to alleviate shaping to a large degree - there may still be shaping present at some stages - it's a necessary component of managing an Uncapped service. The percentage shaping levels are standard across all line speeds though.

So in short we must view this package in it's current state as a streaming/browsing/gaming package rather than a package that can also download? At least until additional bandwidth gets added on your side? As mentioned earlier I am home this December, I don't want to leave Afrihost, but if you can give me/us a honest answer/indication weather we will be less shaped December it would be great. Otherwise my idea was to pause my account on Afrihost and just jump ship for December to someone like openweb or Axxess 10mbps at the current AH 4mbps price. :)
 
Last edited:
Trying to download xampp from sourceforge. Getting horrible speed and disconnects.

The first attempt I got 17mb in about 20 mins then it disconnected. Now ive been downloading 60mins already and only 57/100 mb complete.

I am not shapped, as per the AH clientzone. Any idea why its so slow? Do you know of any local download where I can get XAMPP for Windows?
 
Anyone else in Durban struggling this morning? Connection is so slow it takes ages just to open a page.

Sorry, late reply. I'm having issues this morning as well. Youtube buffers 144p on my 2mb line.

Forget about the congestion at exchange. This trace is bad.

Tracing route to mybb.co.za [213.133.106.183]
over a maximum of 30 hops:

1 <1 ms<1 ms<1 ms 192.168.1.254
2 104 ms 124 ms 143 ms 105-236-1-1-esr-lo.mtnbusiness.co.za [105.236.1.
1]
3 126 ms 103 ms 100 ms ipc-recieve-ng-1a.mtnbusiness.net [41.181.201.14
]
4 201 ms 235 ms 294 ms ng-mse-1--ng-cr-1-a.za.mtnns.net [196.44.27.68]

5 140 ms 90 ms 92 ms jh-cr-1.za--ng-cr-1.za.mtnns.net [196.44.31.176]

6 * * * Request timed out.
7 318 ms 245 ms * qux--rb-dca-2.za-b.za.mtnbusiness.net [41.181.16
5.119]
8 * * * Request timed out.
9 * * * Request timed out.
10 348 ms 471 ms 512 ms ls-cr-2.uk--tb-cr-1.za.mtnns.net [196.44.31.113]

11 253 ms 265 ms 299 ms am-cr-1.nl--ls-cr-2.uk-a.mtnns.net [196.44.31.18
3]
12 * 273 ms 331 ms am-tpr-1.nl--am-cr-1.nl-a.mtn.net [209.212.111.1
41]
13 412 ms 497 ms 378 ms r1ams1.core.init7.net [195.69.144.210]
14 * * 293 ms r1nue1.core.init7.net [77.109.140.26]
15 397 ms 405 ms 368 ms gw-hetzner.init7.net [77.109.135.102]
16 439 ms 497 ms 372 ms core12.hetzner.de [213.239.245.25]
17 334 ms 304 ms 297 ms juniper3.rz2.hetzner.de [213.239.245.70]
18 390 ms 418 ms 296 ms hos-tr4.ms-ex3k2.rz1.hetzner.de [213.239.193.243
]
19 346 ms * 380 ms dedi183.nur4.host-h.net [213.133.106.183]

Trace complete.


******************************
Shaping Policy: No Shaping Applied
 
So in short we must view this package in it's current state as a streaming/browsing/gaming package rather than a package that can also download? At least until additional bandwidth gets added on your side? As mentioned earlier I am home this December, I don't want to leave Afrihost, but if you can give me/us a honest answer/indication weather we will be less shaped December it would be great. Otherwise my idea was to pause my account on Afrihost and just jump ship for December to someone like openweb or Axxess 10mbps at the current AH 4mbps price. :)

I don't think that's the impression we want to give on our accounts :)
Shaping is being applied a bit heavier than what we'd like to see - however changes are in the pipeline to ease some of the shaping by bringing additional capacity online.

I don't think we can ever give any guarantees around whether or not we'll have shaping - it's something that goes hand in hand with an Uncapped service and will most likely always be there - how heavily or lightly the shaping is being applied is a different story.
 
Trying to download xampp from sourceforge. Getting horrible speed and disconnects.

The first attempt I got 17mb in about 20 mins then it disconnected. Now ive been downloading 60mins already and only 57/100 mb complete.

I am not shapped, as per the AH clientzone. Any idea why its so slow? Do you know of any local download where I can get XAMPP for Windows?

If you aren't being shaped then you shouldn't see slow speed here at all :( Have you tested any other services to see if these are slow too?
 
Sorry, late reply. I'm having issues this morning as well. Youtube buffers 144p on my 2mb line.

Forget about the congestion at exchange. This trace is bad.

Tracing route to mybb.co.za [213.133.106.183]
over a maximum of 30 hops:

1 <1 ms<1 ms<1 ms 192.168.1.254
2 104 ms 124 ms 143 ms 105-236-1-1-esr-lo.mtnbusiness.co.za [105.236.1.
1]
3 126 ms 103 ms 100 ms ipc-recieve-ng-1a.mtnbusiness.net [41.181.201.14
]
4 201 ms 235 ms 294 ms ng-mse-1--ng-cr-1-a.za.mtnns.net [196.44.27.68]

5 140 ms 90 ms 92 ms jh-cr-1.za--ng-cr-1.za.mtnns.net [196.44.31.176]

6 * * * Request timed out.
7 318 ms 245 ms * qux--rb-dca-2.za-b.za.mtnbusiness.net [41.181.16
5.119]
8 * * * Request timed out.
9 * * * Request timed out.
10 348 ms 471 ms 512 ms ls-cr-2.uk--tb-cr-1.za.mtnns.net [196.44.31.113]

11 253 ms 265 ms 299 ms am-cr-1.nl--ls-cr-2.uk-a.mtnns.net [196.44.31.18
3]
12 * 273 ms 331 ms am-tpr-1.nl--am-cr-1.nl-a.mtn.net [209.212.111.1
41]
13 412 ms 497 ms 378 ms r1ams1.core.init7.net [195.69.144.210]
14 * * 293 ms r1nue1.core.init7.net [77.109.140.26]
15 397 ms 405 ms 368 ms gw-hetzner.init7.net [77.109.135.102]
16 439 ms 497 ms 372 ms core12.hetzner.de [213.239.245.25]
17 334 ms 304 ms 297 ms juniper3.rz2.hetzner.de [213.239.245.70]
18 390 ms 418 ms 296 ms hos-tr4.ms-ex3k2.rz1.hetzner.de [213.239.193.243
]
19 346 ms * 380 ms dedi183.nur4.host-h.net [213.133.106.183]

Trace complete.


******************************
Shaping Policy: No Shaping Applied

Are you running any network activity when running the traceroute? This looks like fairly consistent exchange congestion which is causing the slow throughput and poor latency on your end :(
 
Are you running any network activity when running the traceroute? This looks like fairly consistent exchange congestion which is causing the slow throughput and poor latency on your end :(

Nope, nothing else is running.
I recreated my port and voila. This is the second time the last week that I had to recreate port. Something I never had to do with M-Web(cough cough).

Seems that we are back to regular port recreations like Openweb.


C:\Users\Pieter>tracert mybb.co.za

Tracing route to mybb.co.za [213.133.106.183]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 192.168.1.254
2 8 ms 8 ms 9 ms 105-236-1-1-esr-lo.mtnbusiness.co.za [105.236.1.
1]
3 17 ms 17 ms 35 ms ipc-recieve-ng-1a.mtnbusiness.net [41.181.201.14
]
4 18 ms 18 ms 19 ms ng-mse-1--ng-cr-1-a.za.mtnns.net [196.44.27.68]

5 41 ms 24 ms 18 ms jh-cr-1.za--ng-cr-1.za.mtnns.net [196.44.31.176]

6 22 ms 22 ms 22 ms rb-cr-2.za--jh-cr-1.za.mtnns.net [196.44.0.43]
7 17 ms 17 ms 17 ms qux--rb-dca-2.za-b.za.mtnbusiness.net [41.181.16
5.119]
8 43 ms 41 ms 42 ms el-mse-2.za--pe-cr-1.za.mtnns.net [196.44.31.67]

9 40 ms 42 ms 40 ms tb-cr-1.za--tb-cr-2.za-a.mtnns.net [196.44.31.4]

10 194 ms 196 ms 228 ms ls-cr-2.uk--tb-cr-1.za.mtnns.net [196.44.31.113]

11 199 ms 196 ms 196 ms am-cr-1.nl--ls-cr-2.uk-a.mtnns.net [196.44.31.18
3]
12 192 ms 191 ms 191 ms am-tpr-1.nl--am-cr-1.nl-a.mtn.net [209.212.111.1
41]
13 201 ms 199 ms 199 ms r1ams1.core.init7.net [195.69.144.210]
14 217 ms 224 ms 224 ms r1nue1.core.init7.net [77.109.140.26]
15 216 ms 216 ms 215 ms gw-hetzner.init7.net [77.109.135.102]
16 215 ms 214 ms 215 ms core12.hetzner.de [213.239.245.25]
17 212 ms 211 ms 211 ms juniper3.rz2.hetzner.de [213.239.245.70]
18 212 ms 212 ms 214 ms hos-tr4.ms-ex3k2.rz1.hetzner.de [213.239.193.243
]
19 216 ms 212 ms 211 ms dedi183.nur4.host-h.net [213.133.106.183]

Trace complete.
 
Afri*, any news on Afrihosts approach to Telkoms Line Speed Upgrades (specifically lines managed by Afrihost)?
 
Afrihost is awesome, Have been using their capped and uncapped products for years now and also good service.
(no way affiliated with them)
 
Nope, nothing else is running.
I recreated my port and voila. This is the second time the last week that I had to recreate port. Something I never had to do with M-Web(cough cough).

Seems that we are back to regular port recreations like Openweb.


C:\Users\Pieter>tracert mybb.co.za

Tracing route to mybb.co.za [213.133.106.183]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 192.168.1.254
2 8 ms 8 ms 9 ms 105-236-1-1-esr-lo.mtnbusiness.co.za [105.236.1.
1]
3 17 ms 17 ms 35 ms ipc-recieve-ng-1a.mtnbusiness.net [41.181.201.14
]
4 18 ms 18 ms 19 ms ng-mse-1--ng-cr-1-a.za.mtnns.net [196.44.27.68]

5 41 ms 24 ms 18 ms jh-cr-1.za--ng-cr-1.za.mtnns.net [196.44.31.176]

6 22 ms 22 ms 22 ms rb-cr-2.za--jh-cr-1.za.mtnns.net [196.44.0.43]
7 17 ms 17 ms 17 ms qux--rb-dca-2.za-b.za.mtnbusiness.net [41.181.16
5.119]
8 43 ms 41 ms 42 ms el-mse-2.za--pe-cr-1.za.mtnns.net [196.44.31.67]

9 40 ms 42 ms 40 ms tb-cr-1.za--tb-cr-2.za-a.mtnns.net [196.44.31.4]

10 194 ms 196 ms 228 ms ls-cr-2.uk--tb-cr-1.za.mtnns.net [196.44.31.113]

11 199 ms 196 ms 196 ms am-cr-1.nl--ls-cr-2.uk-a.mtnns.net [196.44.31.18
3]
12 192 ms 191 ms 191 ms am-tpr-1.nl--am-cr-1.nl-a.mtn.net [209.212.111.1
41]
13 201 ms 199 ms 199 ms r1ams1.core.init7.net [195.69.144.210]
14 217 ms 224 ms 224 ms r1nue1.core.init7.net [77.109.140.26]
15 216 ms 216 ms 215 ms gw-hetzner.init7.net [77.109.135.102]
16 215 ms 214 ms 215 ms core12.hetzner.de [213.239.245.25]
17 212 ms 211 ms 211 ms juniper3.rz2.hetzner.de [213.239.245.70]
18 212 ms 212 ms 214 ms hos-tr4.ms-ex3k2.rz1.hetzner.de [213.239.193.243
]
19 216 ms 212 ms 211 ms dedi183.nur4.host-h.net [213.133.106.183]

Trace complete.

The results are looking much better!
Port errors aren't something that come from an ISP I'm afraid :(

A fault may need to be logged if you notice this issue occurring more often.
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X