Flatrate Schmatrate

[pip]

Why not try to respond?

Methinks he does not understand the points I make! LOL.

 

[rwenzori]

Water itself doesn't cost anything.

Purified water costs.

 

[Peter_J]

As mentioned above, 'bandwidth' is not a physical product.

Agreed!. That is why we should pay for line speed not data throughput. If you connect to any of the big backbones in the USA you pay for the rated speed of your connection, not what you transfer. It is just our beloved Telkom who makes you pay for both.

 

[rwenzori]

It is just our beloved Telkom who makes you pay for both.

Judging from his posts I figure this Henk must work for them

 

[jabulani]

Electricity, water or bandwidth are very much the same in concept.

I disagree. As Pip says, there is a significant marginal cost to the next kilowatt or kilolitre, and virtually zero to the next megabyte. A fundamental difference.

 

[Robin Hood]

600 rand should include a 4 meg line and atleast 50 gig traffic and uncapped local

I agree 100%

 

[henkk78]

You know really, I can assume only that you are now just trolling for fun. Or that your understanding of English is sub-standard. Go find out what "impute" means. I have given you plenty counter arguments that you, like most trolls, ignore. Do you understand the principle of marginal cost? No, I figured not. If I am behaving like a 3-year old it is only because of vainly attempting to communicate in terms you might understand.

You're quite right Pip, I did not know what "impute" meant before I read your post, so I looked it up.

Yes, I do understand the principle of marginal cost. Note, I'm not saying necessarily that bandwidth DOES have marginal cost, all I said is that it's similar in concept to water.

Now, IF you believe that water DOES have marginal cost, and network traffic DOES NOT:

Could you PLEASE focus on proving that it doesn't have a marginal cost, rather than degenerating into calling people trolls... I mean, come on, seriously... It can't be that difficult for you to assemble a reasonable argument.

Let me lay it out for you again from my side:

1. Bandwidth is dependant on infrastructure (technically, capacity & bandwidth, let's differentiate for the sake of argument)
2. The more capacity needed, the more infrastructure.
3. The more infrastructure, the more costs.

I compare this to water, because water also doesn't really cost more just because you're using more. Whether I use 10 litres, or a 1000 litres, the pipes are there and gravity is used to move it down from the reservoir to my house.

 

[alma-tadema]

Same could be said for electricity. It's just energy 'conversion'.

Likewise, water could be seen as an infinite product as well.

You don't make sense, the only cost with bandwidth is the "pipes" and the maintenance of those "pipes" it costs the same if you use it or not... not even water or electricity comes near to bandwidth which is a service not a product.

 

[rwenzori]

I compare this to water, because water also doesn't really cost more just because you're using more. Whether I use 10 litres, or a 1000 litres, the pipes are there and gravity is used to move it down from the reservoir to my house.

LOL. You and captainwifi should get together!

I suppose the Good L0rd just supplied the water pre-chlorinated out of heaven into your reservoir!

 

[alma-tadema]

I agree with someone that compared bandwidth to a facility.

Basicly it would be uneconomical and extremely wastefull to set aside bandwidth for every user for his or her exclusive use. You only do it where it is required and physically necessary like the last mile, that is also why the last mile is so expensive. Network circuits cost the same if it is used 1 % or 100 % of the time, you want too extract the most value out of your spent costs by trying too use it as much as possible. That makes the circuit model totally uneconomical and extremely wastefull. So you run time slices on the circuit, every user get a slice of time to use on the circuit, now what happens when a user doesn't need to use his allocated share, it just skips aheah until it finds a user and will give one user sole use of the circuit if no one elses uses it. Now that user cannot hog the circuit because when another users comes around he just claim his share. This is obviously not unfair after all you want to use the already payed for circuit to the maximum. This is called statistical multiplexing in network speak, and is the fundamental concept behind packet switched networks. Now what happens is that most people will see fantastic speeds most of the time because network traffic is very bursty, even when there are huge bottlenecks. This result in ISP selling broadband when it is really "broadband most of the time".
Now when it comes too selling the service too the consumer while it is very efficient from the sellers perspective, the low users is essentially susidizing the power users if there is a flat fee. So why don't we charge per "time slice" rather than shared access, that is just fair isn't it? The reason this isn't done is because it result in inefficiency which raises prices for the seller, light and power user at the end of the day. The ultra light users end up paying so little, that they may become uneconomical too have as a customer, while the power user seriously restrict their usage after the shock of their first bill resulting in an idle circuit that should be making money. So in the end of the day you still have some users subsidizing others but with the added ineficiency thrown in this time.

Another justification is that power user degrade the experience of the light users at peak times, after all the light user should get a very good service at the few times they are actually using it. This is easily solved by allowing those that for example transferred 100GB this month, only a "time slice" every second round. This is called weighted round-robin. In real life it isn't as sophisticated as that usually they cap your speed or cut you off during peak times.

There is a reason that those countries with flat fee models have cheap and high internet and broadband penetration.

 

[henkk78]

LOL. You and captainwifi should get together!

I suppose the Good L0rd just supplied the water pre-chlorinated out of heaven into your reservoir!

Yes, that's a good point! Moving water doesn't JUST consist of it falling out of heaven, it needs cleaning.

In the same way, the internet doesn't JUST consist of a bunch of copper wires connected to each other, it needs management. The more movement of data, the more management required and the higher the costs.

 

[.geek]

Proof of this? Since when does transferring more data require more "management"?

 

[alma-tadema]

Yes, that's a good point! Moving water doesn't JUST consist of it falling out of heaven, it needs cleaning.

In the same way, the internet doesn't JUST consist of a bunch of copper wires connected to each other, it needs management. The more movement of data, the more management required and the higher the costs.

Yes more data means more management, the problem is that it is machine management, do your computer today really cost more and do the same in your world? In that case move here immediately! Costs per instruction and bit/s has dropped like a stone the last decade, a broadband connection cost less than a single phone 15 years ago in the developed world.
The only human management is of the physical infrastructure, that is only growing slowly and the costs is covered by the new customer base not the existing users. There is no direct human labor involved in delivering bits, otherwise the cost will have exploded exponentially in the telecoms sector.

 

[rwenzori]

In the same way, the internet doesn't JUST consist of a bunch of copper wires connected to each other, it needs management. The more movement of data, the more management required and the higher the costs.

Only if your name is Telkom and you have to "manage" things like P2P to a trickle. Even then how does the volume of data cost more? You have routers purchased and configured to handle the line speeds, whether you have implemented Qos or whatever or not. Once the infrastructure appropriate to the line speeds is there the volume of gigabytes pumped through is immaterial. Whereas the volume of purified water pumped through your tap is a material cost.

If you still think more data costs more to manage please explain where and how, given an infrastructure set up for desired line speeds.

 

[alma-tadema]

Proof of this? Since when does transferring more data require more "management"?

Yep, he thinks periodic upgrades to provide bandwidth is some horrendous continious management exercise. It is a once off sunk cost with maintenance cost, use it to its max or let it lie idle it is going too cost the same. Yes, you have to do it frequently to maintain competitive, but that is the telecom market and newer equipment is alot faster, better and in some cases cheaper, rarely is it more expensive unless it has some greater than normal added value.

 

[pip]

If you still think more data costs more to manage please explain where and how, given an infrastructure set up for desired line speeds.

Well, if you hire a 64k diginet between your two offices from Henkkom, you must pay US$36/Gb for Henkkom techs to manually open up and check your IP packets, to trash those containg porn, to find bottlenecks and manually massage your data through them, to go and unstick your kreepy-krauly, and to manually go reset all your IPs at midnight. Costs money pal!!!

 

[ToSsMaStR]

Consider the recent upgrading of 1 mb lines to 4 mb, it still is the same line is it not. I dont think telkom had to rip out everyones lines and exchange and replace it with a better one? did they?

But has n't it cost more money to cope with the huge amount of data going through, those same lines?

- Then again i could be wrong?

 

[pip]

But has n't it cost more money to cope with the huge amount of data going through, those same lines?

No-one is denying that infrastructure costs money. The debate is around whether, once the infrastructure is in place, more data throughput costs more or the same.

 

[bwana]

No-one is denying that infrastructure costs money. The debate is around whether, once the infrastructure is in place, more data throughput costs more or the same.

iIMO it costr more.

Contention ratio's on adsl it means that x number of people can share a line. With too many ppl running flat out then either usability degrades to a point where they lose customers or more lines are required.

After all, the amount of data that can be fed through a pipe is finite.

 

[pip]

iIMO it costr more.

Contention ratio's on adsl it means that x number of people can share a line. With too many ppl running flat out then either usability degrades to a point where they lose customers or more lines are required.

After all, the amount of data that can be fed through a pipe is finite.

With respect I think you both make the point and miss the point, even with contention ratios. If you are going to be say an ISP, you will calculate that you will contend say 10 384k line users ( or whatever - I am not an ISP ) per 1 mbit pipe you hire from a top level ISP. Your cost for the line ( in addition to known and set costs for routers etc ) per user will be cost-of-1mb-line/10 ( unless that ISP is Telkom of course ). If they are all old grannies who only send email once a day, or if they are rabid porno leechers, that cost does not vary. If you get more users, you must hire more infrastructure - another line. If all your users want to go to 512 from 384, you need more line capacity so they must pay more, whether they pump 1 kb or 1Gb per day.