• You are losing out on amazing benefits because you are not a member. Join for free. Register now.
  • Big Two-Day Giveaway - Win an Amazon Kindle, a Mystery Gadget and Branded Gear. Enter Here.
  • Test your broadband speed and win prizes worth R5,000. Enter here.

How South Africans will pick their fibre speeds online and only pay for what they use

Newsfeed

MyBroadband Newsfeed
Joined
Jun 28, 2017
Messages
5,856
#1
How South Africans will pick their fibre speeds online and only pay for what they use

The rapid pace of fibre growth and increasing competition in the fibre environment will lead to drastically different customer packages than what we see today.

This is according to Dark Fibre Africa (DFA) acting chief strategy officer Vino Govender, who told MyBroadband the way consumers pay for fibre packages will be very different five years from now.
 

justplain

Expert Member
Joined
May 14, 2004
Messages
1,127
#2
The way we pay? Does he mean HOW much we'll pay vs speed? Content providers (Netflix, HBO etc..) are not really cost prohibitive...it's the network access and speed that hold the biggest cost.
 

Johnatan56

Honorary Master
Joined
Aug 23, 2013
Messages
20,700
#3
“Using a system like this, as an end user I could change my bandwidth from 20Mbps to 100Mbps for the two hours, and the system will configure the network automatically and configure your billing accordingly for that period,” said Govender.
This sounds pretty dumb as setting up the link to be able to do 100Mbps burst for a random two hours every week is near the same as setting it to be a permanent 100Mbps due to client's use-case, or am I missing something?
 

Tinuva

The Magician
Joined
Feb 10, 2005
Messages
7,990
#4
This sounds pretty dumb as setting up the link to be able to do 100Mbps burst for a random two hours every week is near the same as setting it to be a permanent 100Mbps due to client's use-case, or am I missing something?
Naw. It sounds like the same type of billing and use as on cloud based systems, where this is the connection speed vs running a big instance for 2 hours then changing it back down to a small instance for most of the month. Scheduled scaling of your line speed.
Cool idea, not sure if it will really work out, even when the systems are that automated.

Also, I am fairly certain Vumatel can already do this, just its easier to leave it at one speed for billing.
 

Johnatan56

Honorary Master
Joined
Aug 23, 2013
Messages
20,700
#5
Naw. It sounds like the same type of billing and use as on cloud based systems, where this is the connection speed vs running a big instance for 2 hours then changing it back down to a small instance for most of the month. Scheduled scaling of your line speed.
Cool idea, not sure if it will really work out, even when the systems are that automated.

Also, I am fairly certain Vumatel can already do this, just its easier to leave it at one speed for billing.
Yeah, but those cloud providers can scale it up/down with no significant change in hardware outside of their servers. For the last mile, all the switches etc. will have to be built to accommodate it, it's a better business model to allow for customers to have access to a permanent 100Mbps with 5% overall utilization, than charging less and having it be 20/40/100 as needed.

Only side I see this work out all right is for the interest side, but with peering agreements/more CND/caching servers that's not as impactful anymore (e.g. most of steam is hosted in CT for CT people).

On an e.g. Cloud Server you had to previously reserve that amount of memory etc. for the VM, having fiber + ISP is the same due to having to build it all beforehand. It's not like micro services/serverless where you can cheaply spin up a fiber connection and destroy it with near no costs.

What Vuma charges is probably base + maximum speed that can be utilized to pay for upgraded switches etc.
 

Tinuva

The Magician
Joined
Feb 10, 2005
Messages
7,990
#6
Yeah, but those cloud providers can scale it up/down with no significant change in hardware outside of their servers. For the last mile, all the switches etc. will have to be built to accommodate it, it's a better business model to allow for customers to have access to a permanent 100Mbps with 5% overall utilization, than charging less and having it be 20/40/100 as needed.

Only side I see this work out all right is for the interest side, but with peering agreements/more CND/caching servers that's not as impactful anymore (e.g. most of steam is hosted in CT for CT people).

On an e.g. Cloud Server you had to previously reserve that amount of memory etc. for the VM, having fiber + ISP is the same due to having to build it all beforehand. It's not like micro services/serverless where you can cheaply spin up a fiber connection and destroy it with near no costs.

What Vuma charges is probably base + maximum speed that can be utilized to pay for upgraded switches etc.
Its kind of exactly the same. Cloud providers do need the hardware around for you to spin instances up and down. Same goes for the fibre provider, they will need the hardware and backhaul, but allow you to spin your connection speed up/down.
 

Johnatan56

Honorary Master
Joined
Aug 23, 2013
Messages
20,700
#7
Its kind of exactly the same. Cloud providers do need the hardware around for you to spin instances up and down. Same goes for the fibre provider, they will need the hardware and backhaul, but allow you to spin your connection speed up/down.
No, because the cloud providers can use that same hardware across everyone, you can have user X in Durban spin up an instance, terminate it, then have user Y use that same hardware.

For a last mile fiber provider, you have to roll out the probably most expensive 2/3km at the last part, this merges into more users upstream, merges again, etc. but the max capacity must always be there. 5% utilization costs the same as 90% if on the already built network. For the cloud server, you can rent out the performance and swap quickly, scaling up/down the number of instances, reducing costs.
 

HavocXphere

Honorary Master
Joined
Oct 19, 2007
Messages
31,487
#8
“By that I mean you could, as an end user, access a portal and choose the amount of speed you want for a certain amount of time, on the fly,” said Govender.
Nobody wants that sht. It's like that X gig night time data after hours Y if it's a full moon, but not on a thursday and you need to click on a portal button for the speed boost.

If you can't provide uncapped at a reasonable price: Telkom had the right idea with their softcaps...give people a lot of data and then throttle for excessive usage.
 

TheRoDent

Expert Member
Joined
Aug 6, 2003
Messages
3,006
#9
This "Turbo Button" bullsh*t has been touted since forever by Ericsson (now fscked) and Nokia (now fscked) and it has never taken off.

But nice try for trying to pull a dead cow "out of the sloot" Mr Govender.

“Bandwidth is becoming so much more available, who knows whether we will even have capped and uncapped? Maybe it might be just uncapped in future.”
The future is here already Mr Govender. Just signup with Cool Ideas. :)
 
Last edited:

Tinuva

The Magician
Joined
Feb 10, 2005
Messages
7,990
#10
No, because the cloud providers can use that same hardware across everyone, you can have user X in Durban spin up an instance, terminate it, then have user Y use that same hardware.

For a last mile fiber provider, you have to roll out the probably most expensive 2/3km at the last part, this merges into more users upstream, merges again, etc. but the max capacity must always be there. 5% utilization costs the same as 90% if on the already built network. For the cloud server, you can rent out the performance and swap quickly, scaling up/down the number of instances, reducing costs.
I agree with your reasoning that cloud providers have a much larger economies of scale and that count in their favor.

But for FTTH its different. The network equipment will be there regardless of if everyone is on 100 or 20 or 10.
What you will find is, that most customers will work out, they dont need 100 or 200 or 1000 lines, but will settle for the 4 or 10 or 20 lines. The equipment is still there, it can do more for every house. The only exception to this would be IPC, but luckily only Openserv is limited to that model.

So what I see in this article, is that ISPs can make extra money, by allowing those customers that mostly only can afford 4,10 or 20 for the whole month, to up their speed for a few hours, for whatever reason, and only add a marginal cost to the customer. If many customers over the country then do this, thats extra money recouped by the ISP/Fibre provider. Remember, the equipment and lines are already installed, all this is, is a software config change on the port, that will now be automated on both the hardware config and billing side. And this is why I compared it to the cloud.

I see the business case in this article, if you don't thats fine. I am only thinking in terms of uncapped packages here, with capped being irrelevant at this stage.
 

GhostSixFour

Mafia Addict
Joined
Nov 9, 2009
Messages
11,394
#11
If I can set this on an automated schedule, great. However, I suspect that the pricing model will be skewed towards peak time - i.e. you pay more between 18:00-22:00 - so probably even leaving it on 1mbps whilst at work for the day, probably won't recoup any pricing difference for me vs just having the line set at 20.

If they're hoping I subscribe to better speeds cause my internet is slow - fsck em. The only reason it'll be slower is if they set it slower. 20mbps is sufficient for what I want - and working currently.
 

kianm

Executive Member
Joined
Jan 13, 2014
Messages
9,938
#12
I was here thinking it is a great idea but i see many disagree, i still think there is a place for dynamic services though
 
Top