can someone explain the uncapped cost for the ISP

[Lightscribe]

Should ISP's not be focusing then on fibre lines to large complexes etc?

Telkom's strategy is data centres and gated communities/complexes. Dig around their site.

It's just another reason that we chose the IS Uncapped Basic product to resell. IS are huge, they are setup and have been doing this a very long time. The odds are completley in their favour that they will come out with the better service...

Naspers has deep pockets too. The two of them, together, going at this for the next few years, makes me smile. Bye bye unsustainable...Hello new world!

 

[Abe]

South Africa's issue is that the vast majority of people have a 384k line and download speed is very slow so a LOT of people are always 'using' the internet. Therefore contension ratios have to be a LOT less for the least profitable line speed.... which doesn't make economic sense.

The more people we can get on 4M+ the more likely the 50:1+ contention ratios can be maintained and actually give every 100% line-speed service all the time.

I was just trying to get my head around this part. Would it not be beneficial to rather have people with slower connections as you can sell more accounts into the same pipe?

 

[nakedpeanut]

I was just trying to get my head around this part. Would it not be beneficial to rather have people with slower connections as you can sell more accounts into the same pipe?

ppedrick kinda answered that:

South Africa's issue is that the vast majority of people have a 384k line and download speed is very slow so a LOT of people are always 'using' the internet. Therefore contension ratios have to be a LOT less for the least profitable line speed.... which doesn't make economic sense.

Must say +1 to ppedrick.. Explained very well!!

So basically from my understanding of the current situation, people like AfriHost cant offer as viable an uncapped solution as mweb, as I assume AfriHost's user base is not filled with "moms and pops" like mwebs is.
Thus their power users arent getting subsidized buy as large a population. So the 80-20 rule is a little skewed in their case? Hence the 30, 60, 90, 120GB limiting they have applied to their accounts?

In essence which ISP would be able to offer an better more sustainable uncapped offering in our opinion? @ppedrick

Currently with AH and contemplating a move to mweb

 

[Gordon_R]

As they already had 150,000 ADSL users before the uncapped offering you'd think their network would initially be a minimum of 2-3GBit... but you never know. They could have bought just 1Gbit or 10Gbit. That will be the bit that decides on if they are good or bad... and alo the fact they are now buying SAT-3 for their redundancy... although how much is yet to be seen/known.

I think MWeb bought quite a large pipe on SEACOM. The reason I sugest this is that tracert to many of the big sites goes via SEACOM, and many users report that YouTube vidoes are fast ( a big bandwidth hog). When the connection to SEACOM was erratic on Thurs 1st , international traffic seemed to grind to a halt, even with SAT-3 backup.

 

[ppedrick]

So basically from my understanding of the current situation, people like AfriHost cant offer as viable an uncapped solution as mweb, as I assume AfriHost's user base is not filled with "moms and pops" like mwebs is. Thus their power users arent getting subsidized buy as large a population. So the 80-20 rule is a little skewed in their case? Hence the 30, 60, 90, 120GB limiting they have applied to their accounts?
In essence which ISP would be able to offer an better more sustainable uncapped offering in our opinion? @ppedrick
Currently with AH and contemplating a move to mweb

The problem that AH have is they have chosen to resell an IS product where theybuy bandwidth 'per GB'.
Threfore, AH can no afford to have people download 100% of the time as they literally do pay for it!
So the system they use can 'throttle' and 'shape' to 5 different levels dependent on the user's usage... it's classed as uncapped because they will not cut you off... just throttle you down to 128kpbs.

The other cheap IS option is the IS Uncapped Basic which is shaped and managed by the Tier 1 provider themselves.
This is the one our company and many others have chosen to resell as we believe that it is the responsibility of the Tier 1 to provide the best service... not us! Therefore, we have come out in the open and are trying to promote complete reseller transparency so the customer knows what they are buying. Also, so that we can all pile on the pressure to the Tier 1 to constantly improve their performance.

 

[nakedpeanut]

The problem that AH have is they have chosen to resell an IS product where theybuy bandwidth 'per GB'.
Threfore, AH can no afford to have people download 100% of the time as they literally do pay for it!
So the system they use can 'throttle' and 'shape' to 5 different levels dependent on the user's usage... it's classed as uncapped because they will not cut you off... just throttle you down to 128kpbs.

The other cheap IS option is the IS Uncapped Basic which is shaped and managed by the Tier 1 provider themselves.
This is the one our company and many others have chosen to resell as we believe that it is the responsibility of the Tier 1 to provide the best service... not us! Therefore, we have come out in the open and are trying to promote complete reseller transparency so the customer knows what they are buying. Also, so that we can all pile on the pressure to the Tier 1 to constantly improve their performance.

I see, thank you.. Ya it makes more sense to me from an uncapped perspective to sell IS's product like that than per GB..

 

[Lightscribe]

The problem that AH have is they have chosen to resell an IS product where theybuy bandwidth 'per GB'.
Threfore, AH can no afford to have people download 100% of the time as they literally do pay for it!
So the system they use can 'throttle' and 'shape' to 5 different levels dependent on the user's usage... it's classed as uncapped because they will not cut you off... just throttle you down to 128kpbs.

The other cheap IS option is the IS Uncapped Basic which is shaped and managed by the Tier 1 provider themselves.
This is the one our company and many others have chosen to resell as we believe that it is the responsibility of the Tier 1 to provide the best service... not us! Therefore, we have come out in the open and are trying to promote complete reseller transparency so the customer knows what they are buying. Also, so that we can all pile on the pressure to the Tier 1 to constantly improve their performance.

Someone has pointed out in a thread, that the Afrihost account will allow around 175GB/month, with the way that it's throttled. That's perfect for 384 and 512K users, who will not even get to the 175GB.

What Afrihost has done is merely to prevent the really heavy users/4/8/10meg line speeds, from using more than 175 gigs per month. Totally fair.

I'm sure the IS basic product will be great for many also, since IS will make sure their products perform as well as, if not better than the MWEB ones.

For me, (since I can finally get 4meg due to recent upgrades) the issue now, is finding the service with the lowest latency and highest, most consistent, package speed.

 

[Abe]

I think MWeb bought quite a large pipe on SEACOM. The reason I sugest this is that tracert to many of the big sites goes via SEACOM, and many users report that YouTube vidoes are fast ( a big bandwidth hog). When the connection to SEACOM was erratic on Thurs 1st , international traffic seemed to grind to a halt, even with SAT-3 backup.

That is my concern with both MWeb and IS. They have both bought big chunks on SEACOM but SAT3 is far more expensive. It will be a slightly risky endeavor until EASSY lands when there will be a nice compliment for SEACOM.

 

[Gordon_R]

I think MWeb bought quite a large pipe on SEACOM. The reason I sugest this is that tracert to many of the big sites goes via SEACOM, and many users report that YouTube vidoes are fast ( a big bandwidth hog). When the connection to SEACOM was erratic on Thurs 1st , international traffic seemed to grind to a halt, even with SAT-3 backup.

That is my concern with both MWeb and IS. They have both bought big chunks on SEACOM but SAT3 is far more expensive. It will be a slightly risky endeavor until EASSY lands when there will be a nice compliment for SEACOM.

MWeb international traffic (SEACOM) ground to a halt again for about 20 mins this evening (a few hours after I made that comment).

The biggest cost in bandwidth is having 24/7 redundancy, which would be prohibity expensive to service at 100% capacity. It is also expensive to switch traffic instantaneously, since this requires more equipment AFAIK. Obviously priority traffic is given preference when the cheaper route goes offline, and shaping on other traffic will be harsher...

 

[SaiyanZ]

I was just trying to get my head around this part. Would it not be beneficial to rather have people with slower connections as you can sell more accounts into the same pipe?

They are hoping that people will pay R499 for a 4Mb/s line rather than R219 for a 384kb/s line and still use the same amount of GB's. So they make R280 more for the same input costs.

However if people are downloading constantly at full speed the whole day then your argument holds. The 384kb/s R219 account would be more profitable to the ISP as it is around 10 times slower than a 4Mb/s R499 account, but the 4Mb/s account is not 10 times the price.

 

[Gordon_R]

They are hoping that people will pay R499 for a 4Mb/s line rather than R219 for a 384kb/s line and still use the same amount of GB's. So they make R280 more for the same input costs.

However if people are downloading constantly at full speed the whole day then your argument holds. The 384kb/s R219 account would be more profitable to the ISP as it is around 10 times slower than a 4Mb/s R499 account, but the 4Mb/s account is not 10 times the price.

I asked this exact question on the MWeb Official Q&A, but never got an answer. I would like to believe that the 4Mbps package gets 12 times the speed, but this is not specified in any of the Terms & Conditions. The shaping policy is totally opaque, particularly with regards to international traffic:
http://mybroadband.co.za/vb/showthr...cial-Q-amp-A&p=3722929&highlight=#post3722929

This is one of the reasons why I have not yet upgraded to 4Mbps uncapped. (I'm on 4 Mbps capped, and I don't do newsgroups, etc, so its not worth it for me).

 

[nakedpeanut]

I asked this exact question on the MWeb Official Q&A, but never got an answer. I would like to believe that the 4Mbps package gets 12 times the speed, but this is not specified in any of the Terms & Conditions. The shaping policy is totally opaque, particularly with regards to international traffic:
http://mybroadband.co.za/vb/showthr...cial-Q-amp-A&p=3722929&highlight=#post3722929

This is one of the reasons why I have not yet upgraded to 4Mbps uncapped. (I'm on 4 Mbps capped, and I don't do newsgroups, etc, so its not worth it for me).

If you don't do about 17GB a month Assuming R30 per GB and uncapped costs R500, there is no point as you'd be paying more for cap you dont use...

 

[Abe]

MWeb international traffic (SEACOM) ground to a halt again for about 20 mins this evening (a few hours after I made that comment).

The biggest cost in bandwidth is having 24/7 redundancy, which would be prohibity expensive to service at 100% capacity. It is also expensive to switch traffic instantaneously, since this requires more equipment AFAIK. Obviously priority traffic is given preference when the cheaper route goes offline, and shaping on other traffic will be harsher...

Ideally you would want a 50/50 split between the two lines and losing one would make it still bearable for everyone.

 

[nakedpeanut]

Ideally you would want a 50/50 split between the two lines and losing one would make it still bearable for everyone.

Ideally yes... But with the lines having such varied capacity.. Realistically I doubt it...

 

[Lightscribe]

Why not utilise both of the cables' constantly growing amount of pipes leased/purchased, to 100% capacity, thus speeds merely drop somewhat, when one goes down...? That makes most financial sense.

Especially since prices will drop further as more cables land, increasing the redundancy choices each time, meaning that when any one of say, five cables go down, there is only a 20% increase in load on the system.

 

[nakedpeanut]

Why not utilise both of the cables' constantly growing amount of pipes leased/purchased, to 100% capacity, thus speeds merely drop somewhat, when one goes down...? That makes most financial sense.

Especially since prices will drop further as more cables land, increasing the redundancy choices each time, meaning that when any one of say, five cables go down, there is only a 20% increase in load on the system.

I'm sure ISP's try to use as much of the capacity that they have bought.
You are want to make full use of the asset you have purchased.

 

[Lightscribe]

I'm sure ISP's try to use as much of the capacity that they have bought.
You are want to make full use of the asset you have purchased.

Yes, although I meant it as regards the points put forth by Gordon_R.

MWeb international traffic (SEACOM) ground to a halt again for about 20 mins this evening (a few hours after I made that comment).

The biggest cost in bandwidth is having 24/7 redundancy, which would be prohibity expensive to service at 100% capacity. It is also expensive to switch traffic instantaneously, since this requires more equipment AFAIK. Obviously priority traffic is given preference when the cheaper route goes offline, and shaping on other traffic will be harsher...

I would rather have my clients experience a loss in speed, than a loss of connection. That needs to happen just two or three times and you have lost a chunk of clients...Then I would "have to" implement the automatic switching, at the cost that I was trying to avoid, with a shattered reputation, in a particularly fickle market.

 

[Gordon_R]

I would rather have my clients experience a loss in speed, than a loss of connection. That needs to happen just two or three times and you have lost a chunk of clients...Then I would "have to" implement the automatic switching, at the cost that I was trying to avoid, with a shattered reputation, in a particularly fickle market.

I agree, although I would imagine that business clients have a different set of requirements from domestic users.

Most of the breaks seem to have been in the evening (or on a long-weekend), when demand would be different from office-hours.

I think some of this infrastructure and cost/benefit tradeoff is 'uncharted territory', and teething problems will occur...