How useful is ICASA?

In reality implementing the requirement to stop periodic resets will cost a lot of money, and increase Telkom's costs.
In an environment where there is competition, companies would consider costs like these to be standard running costs. Costs required in order to improved their services and enable them to offer better "goods" than their competitors.

Problem again is that until now we have not had that, even so it is still very early with Neotel and we have yet to see if we will have a MTN/Vodacom situation where both end up reaping HUGE profits, without having to give anyone anything back since we all have no options but to accept what is given to us., That is where ICASA should step in to protect us from these situations.

I would like to know how releasing contention ratios would cause telkom an increase in costs? Other than the fact that once they are released it would become known that Telkom are currently way out of "spec" with international norms, and in order too apply to this norms, they would end up losing profits. Loss of profits (gained by unfair practices) is not an increase of costs.

I was briefly browsing through neotels new site today (new for me I've been out of the loop for a few months) and was looking at their planned services. One of the first things I noted was their mention of contention ratio's on some of their services.. If they can do it.. why not telkom?
 
I realise this is a Noob question, but can someone please explain contention ratio's? How does this compare to latency figures or do they have nothing to do with each other? :o
 
Quickest way to explain...

http://www.getonlinebroadband.com/faqs/faq02.html
or
http://en.wikipedia.org/wiki/Contention

Basically its how many people "share" a given service or available bandwidth. So say is 50:1 for home, 20:1 for bussiness. If you have 100:1 though, say on home. You have 100 people using the same BW which is shared between them all. This will in effect cause you to have slower speeds... If they all use the same service at the same time

So telkom pays X and then has 100 people pay Y for X, meaning a 100% profit compared to other international companies
 
Ok I get it thanks Kalvaer, for the help

Another question, who else has the ability to cause a change to the contention ratio? Is it only Telkom at the exchange level (I presume this is the "bottleneck point" so to speak, the number of people connecting at a single exchange, the point at which the ratio could be measured?), or can the ISP's be responsible as well?

It think this is a dumb question because it must be on Telkom’s side but I don’t want to assume, you know assumptions are the mother of all ****ups

*Edit* O yes, what is decent contention ratio 50:1; 100:1 ?
 
Last edited:
I do believe ISP's can also set contention ratios. Biggest problem is telkom wont tell anyone what they are doing behind their closed doors

As to what is Decent.. Lower = Better. Some US sites I checked have ratios of 5:1 and 10:1, business are as low as 2:1 in some area's. I believe BT in the UK uses 20:1 for business and 50:1 for home
 
Ok I recon I feel enlightened enough to give my opinion on contention ratios as well, thanks.

I recon Telkom won’t release the ratio’s cause they know that we are gonna poo ourselves!
Wouldn’t the entire cookie crumble for them though if they did? I see some people are asking why Telkom won’t release the ratios when Neotel does, but if I understand correctly, if they release the ratio figure’s then people will know how many users on a line, and then people would be able to calculate the cost per user (knowing the cost of the line? Am I correct?), and then ICASA can’t use the excuse of waiting for Telkom’s costing info, and then Telkom will be beaten with a wet sock for over charging, and then and then and then… In the end, it’s not in their interest; not because it will cost extra, as some may contend, but because it will show them up for what they are, greedy little pigs!

Am I wrong?
 
I'm not suggesting that it is your job to fight ISPs' battles. ISPs are busy doing that. What I am suggesting is that it is necessary for everyone to be congnisant of the indirect effects of their actions.
Is this fight a silent one?

It's probably just me but I'd expect a more proactive stance, lawsuits, advertising campaigns, complaints to the various commissions.
 
Surely some mistake

This means, for instance, that if Telkom's service delivery becomes more expensive, rightly or wrongly this strengthens Telkom's argument that we as consumers are not being overcharged.

Following this logic any benifits that consumers want in service or pricing will end up costing Telkom money and so we can never expect them?

We only need to look at the dialup Internet access market to see how competition has benefited the consumer, where prices have dropped considerably over the years.

Really? Have dial up prices not always been around the R100/ month mark give or take R20? Overseas I think many people have free dial up accounts?

In reality implementing the requirement to stop periodic resets will cost a lot of money, and increase Telkom's costs
Fair enough and I agree about misdirected effort, however, itll be a nice thing to have later but for now Ill settle for NO TRAFFIC SHAPING, Thatll probably reduce cost in hardware and measuring devices but cost telkom cash as they wont be able to sell a 2 tier luxury/Pauper product so we cant expect that? Local bandwidth not adding to caps, but then Telkom will lose out on revenue
on potential sales so we cant expect that either?

Do people really not think there is such a thing as unreasonable profit? If there is ICASA must get involved.

ICASA may have protected consumers from Telkom charging us R1/second or more on the phone as Im sure they would have tried if they could.

I
 
"Originally Posted by geoffr
I'm not suggesting that it is your job to fight ISPs' battles. ISPs are busy doing that. What I am suggesting is that it is necessary for everyone to be congnisant of the indirect effects of their actions."

Is this fight a silent one?

It's probably just me but I'd expect a more proactive stance, lawsuits, advertising campaigns, complaints to the various commissions.

Thought about, and I recon geoffr just doesnt want you to poke the sleeping Telkom. "if your nice to Telkom then they might be nice to you"
 
Is this fight a silent one?

It's probably just me but I'd expect a more proactive stance, lawsuits, advertising campaigns, complaints to the various commissions.

Unfortunately for the public, a lot of work goes on behind closed doors, and is not immediately visible. Much progress can often be gained through negotiation, rather than through public actions. Even though lawsuits are in the public domain (once they come to court) most people are still not aware of these. Lawsuits also take incredibly long, so lawsuits do not always get results as quickly as negotiated paths often do.
 
So essentially we're asked to believe that a silent fight with no tangible (as yet) benefits is being waged against or perhaps for Telkom?
 
Ok I recon I feel enlightened enough to give my opinion on contention ratios as well, thanks.

I recon Telkom won’t release the ratio’s cause they know that we are gonna poo ourselves!
Wouldn’t the entire cookie crumble for them though if they did? I see some people are asking why Telkom won’t release the ratios when Neotel does, but if I understand correctly, if they release the ratio figure’s then people will know how many users on a line, and then people would be able to calculate the cost per user (knowing the cost of the line? Am I correct?), and then ICASA can’t use the excuse of waiting for Telkom’s costing info, and then Telkom will be beaten with a wet sock for over charging, and then and then and then… In the end, it’s not in their interest; not because it will cost extra, as some may contend, but because it will show them up for what they are, greedy little pigs!

Am I wrong?

You will actually find that the engineers who run Internet access services are tickled pink by the discussion about contention ratios, as they are meaningless. What really counts is packet loss. If a service is obsubscribed a million to one, but there is no packet loss, users will not see a problem. If however a service is oversubscribed 6:1 and there is packet loss the subscribers will experience terrible performance.

I have seen a service which was specified at 10:1 before launch being changed to around 6:1, because at 10:1 oversubscription the service was terrible due to packet loss, and the customers were complaining.

You may, in the end find that an oversubscription ratio of 50:1 is fine in one area of a network, but 5:1 is needed elsewhere.

The use of oversubvscription ratios should be reserved for the domain of those who are producing business models to determine, before launch, whether or not a particular service may be viable or not. They equate to a rule of thumb and nothing more.

You will find in environments where hard SLAs are provided, oversubscription is never mentioned. For instance, providers of MPLS networks quote metrics like packet loss, round trip delay and jitter (which is only relevant for voice and video but no data services).

From an engineering point of view, it is relevant to look at metrics like queue sizes and packet drops. Oversubscription ratios again tell you nothing about the performance of a service.
 
business are as low as 2:1 in some area's. I believe BT in the UK uses 20:1 for business and 50:1 for home


Anyone know whether Telkom's ADSL 'business users' have a different contention than that of other users?

/Edit Or in Geoffr's terms, whether they have fewer packet losses? Edit/
 
Last edited:
@ geoffr



Would you be able to post a link if you have one for the presentation? I would like to get a better understanding of you're argument as there are one or two point I’m struggling with

Like…



If there isn’t a healthy wholesale market (as we are currently sitting with) what else can we do other than expect ICASA to regulate pricing?

This is just a matter of opinion so don't beat me for it, but how closely do you really think cost is linked to retail pricing for Telkom. By all accounts their mark ups are so high that they should easily absorb the costs of periodic resets, amongst other things

Have a look at http://www.ispa.org.za/iweek/2006/presentations/traceycohen_iweek.zip

Some of the interesting material is unfortunately not in the slides.
 
It is therefore very unlikely that we will see a truly liberalized telecoms environment anytime soon, and we may just have to hope that ICASA gets more funding, that it is given true independence and that the few competent senior staff left do not also depart in search of greener pastures.
funding for what? so that all the employees can receive an salary increase to 900k and still do **** all?
 
How useful is ICASA?

Mmmm... about as useful as having a fire in the living room and all you've got is a sieve to carry water to douse the flames... :D
 
Have a look at http://www.ispa.org.za/iweek/2006/presentations/traceycohen_iweek.zip

Some of the interesting material is unfortunately not in the slides.
Just been through the zipped .ppt, but without slide notes the .ppt is somewhat flat, although slide 8 and some other bits are interesting [not surprising], but nothing to explain why ICASA went so mental & backtracked on what the ADSL regs were supposed to have been.

I stick by what I posted earlier about giving ICASA less public money, and force ICASA to run off of fines imposed on licensees that fail to meet license conditions.
 
What really counts is packet loss. If a service is obsubscribed a million to one, but there is no packet loss, users will not see a problem. If however a service is oversubscribed 6:1 and there is packet loss the subscribers will experience terrible performance.
Not entirely true. The idea with the contention ratios (or the oversubscribing as you mentioned) is that in theory not all the users will be using the BW at the same... In theory this works, but its not always the case.

Lets say we have a 10 Mbit/s Line, shared in a 10:1 ratio.
Now if all the users are on at the same time, that would allow each user to have 1 Mbit/s available to them, But then comes in the idea of "when are all of them online at the same time?". The customers are still paying for it, but how do you make it cheaper for the customer.. Increase the Ratios, drop your price by 25% and get more customers on the same line. Customers are happy because its cheaper and you end up getting more money off the same line :rolleyes:

In order to maximise profit it would be better to use a ratio of 20:1 and if by chance all these users are on at the same time, they will only get 0.5 Mbit/s available to them and it will be put down to a technical fault or even blamed on the line on the other end (dont we all know that story to well.)

Push that ratio up even more to say 50:1 on the same line.. and you are greatly increasing the chances of more people being on the same line. In effect slowing each user down drastically to 0.2 Mbit/s..However the SP is still paying for the same amount of BW, but now making much more money out it.

Each situation needs to be monitored and checked. Ratios of 20:1 might work perfect for a 10Mbit/s line in an area like Rosebank, given all are on 1 Mbit lines and the line usage is at say 50% - 75%. But in an area like Poffadder you could push that up to 50:1 or 100:1 where your usage is around 10%


You will find in environments where hard SLAs are provided, oversubscription is never mentioned. For instance, providers of MPLS networks quote metrics like packet loss, round trip delay and jitter (which is only relevant for voice and video but no data services).
I might be mistaken here, but in these situations the customer is given a Guaranteed min line speed. IE: the ratios are set in such a way that even if all the users of the line are on at the same time, the shared BW will be at your minumim.

I still firmly believe that if we were to find out what the Ratios are that telkom use, we would find them grossly out of spec, indicating huge profits at our expense, and also explaining why people in certain areas battle with slow speeds. If ICASA would force a minimum required line speed, then we can forget about the ratios since Telkom would in effect have to pull in line with this to make it work.

Of course.. Then we would all move over to Latency and Packet loss due to bad maintanence from the service provider :confused:
 
Last edited:
Unfortunately for the public, a lot of work goes on behind closed doors, and is not immediately visible. Much progress can often be gained through negotiation, rather than through public actions. Even though lawsuits are in the public domain (once they come to court) most people are still not aware of these. Lawsuits also take incredibly long, so lawsuits do not always get results as quickly as negotiated paths often do.

my 2 cents: Whatever path is taken, Telkom will fight tooth and nail to protect profits, so they will stall wherever possible, even it means maintaining the appearance of negotiating in good faith only to pull the rug out later on and then be forced into a lawsuit... Their profits are so obscene and time is money! They get richer every second :mad:
 
Unfortunately for the public, a lot of work goes on behind closed doors, and is not immediately visible. Much progress can often be gained through negotiation, rather than through public actions. Even though lawsuits are in the public domain (once they come to court) most people are still not aware of these. Lawsuits also take incredibly long, so lawsuits do not always get results as quickly as negotiated paths often do.

And once the ISP's have negotiated their profit margins they are happy. Screw the paying customers. I trust the big ISP's and ISPA as much as Telkom.
 
Top
Sign up to the MyBroadband newsletter
X