NBN Co chief executive Mike Quigley today said his team had opened discussions with Simon Hackett over the Internode managing director’s NBN pricing concerns, but stopped short of saying NBN Co would be able to address what Hackett has previously described as the company’s “insane” pricing model.
The debate about NBN Co’s pricing model as a whole was kicked off in late March when Hackett gave a landmark speech describing the model as “insane” for small internet service providers, warning that none would survive their walk through the “valley of death” transition from the current copper network to the fibre future envisioned by the Federal Government.
Over the succeeding weeks, NBN Co made a number of public attempts to further explain its pricing model, going into great detail about why it had chosen the specific mix, as well as publishing an online calculator that allows users to calculate projected wholesale costs for providing services to its planned network.
However, Hackett also fired back several times at what he saw as “crystal ball gazing” by the company. In a hearing this morning in Sydney of the Federal Parliament’s NBN committee, independent Senator Nick Xenophon noted Hackett had gotten “pretty fired up” over the issue.
Quigley replied that NBN Co had, subsequent to the war of words, had “ongoing discussions with Mr Hackett and his team about their concerns”. “We will be working through this with Mr Hackett and his team,” he said.
The NBN Co chief noted his company’s pricing model was a balance. The model constitutes a mix between the basic per user charge for connecting end user customers to NBN Co’s network (the ‘Access Virtual Circuit’), as well as a charge based on data usage (the ‘Connectivity Virtual Circuit’). Hackett has argued for the balance between the two ‘$x and $y’ charges to be changed, with the CVC pricing to come down and the AVC pricing to rise.
Quigley said the tranche of smaller operators envisaged under the NBN tended to focus on smaller geographical locations, rather than trying to cover the entirety of Australia. “Their investments will be much more modest,” he said, as a consequence.
In addition, Quigley noted that some aspects of NBN Co’s pricing model were driven by the Australian Competition and Consumer Commission’s decision to set the numbers of points where ISPs will be able to interconnect with the NBN fibre at 121 nationally — rather than 14 as NBN Co itself had favoured. Hackett believes the pricing model advantages large players with existing infrastructure such as Telstra and Optus.
“That’s not our decision — it’s an ACCC decision,” said Quigley.
However, the NBN Co chief rejected comments quoted by Xenophon from NEXTDC chief Bevan Slattery that the CVC pricing model would bring down the NBN. “That’s not what we’re hearing from customers who we’ve done deep dives with,” said Quigley.
In addition, he noted that ISPs themselves would have to have some skin in the game. “It’s also clear that we tested this with quite a number of [retail service providers],” Quigley said. “Would they like our prices to be lower? Of course.”
“At some point an RSP has to make an investment.”
Image credit: NBN Co
Good to see the ACCC are medling via the same sloth-like, reactive, ambulance-at-cliff approach as the current CAN.
Because that has worked out swimmingly so far.
*In addition, Quigley noted that some aspects of NBN Co’s pricing model were driven by the Australian Competition and Consumer Commission’s decision to set the numbers of points where ISPs will be able to interconnect with the NBN fibre at 121 nationally — rather than 14 as NBN Co itself had favoured. Hackett believes the pricing model advantages large players with existing infrastructure such as Telstra and Optus. That’s not our decision — it’s an ACCC decision,” said Quigley.*
well, this just doesn’t make any sense at all.
is NBN Co. suggesting that it would be cheaper to build a national fibre network with 14 POIs as opposed to 121 POIs? i would’ve thought a 14 POI model would involve more investment in backhaul by NBN Co. (which is why some industry players are suprised that CVC charges are as high as they are given the final ACCC determination of 121 POIs).
the fact is NBN Co. needs to earn $Xbln in revenue in order to meet its financial objectives. if you reduce the number of POIs, or whatever abstract chargeable units you choose to devise, you’ll have to increase the charges in a different dimension on a proportionate basis to maintain that same revenue target.
also, the industry complaints do not centre on the structure of the charges (AVC + CVC), but the level of the CVC charges in specific. and the level of CVC charges is clearly set in order to recoup sufficient revenue to pay off the exorbitant cost of pushing fibre to (almost) every household.
even if the ACCC approved a 14 POI model, NBN Co. would still adopt an AVC + CVC pricing structure. furthermore, the higher backhaul investment would necessitate higher CVC charges than currently proposed.
unless NBN Co. has been misquoted in this article, sounds like more dissembling and buck-passing to me.
No they aren’t suggesting this, what will change is the ability for providers to actually provision services easier. It isn’t about price, it’s actually able enabling competition.
Under the current model if you try and run a national RSP you need to buy CVC for every single POI. Let us assume the CVC doesn’t change but the POIs do (which makes sense, you’re not changing the capacity, only where it is accessed). Now let us say you are offering 100Mbps services, under this you will need to purchase large amounts of “redundant” CVC. If plan to provide a service with 20:1 (100Mbps), for example, you need at 40 customers per POI to break the point when you go over the “initial” 200Mbps you will need to provision.
With a higher number of POIs what this results in a higher costs of entry. An ISP like Telstra already has the customers it needs to ensure that they don’t have to buy redundant CVC, but if a new ISP wants to enter the block they are going to have a period where they have to pay more than they are earning in order to provide the advertised services, or only provide to select POIs and expand as they get the revenue.
This was the premise of Simon Hackett’s arguement, he further went on to state that if you want to reduce this problem and stay with a higher number POIs you need to rebalance the CVC and AVC charges such that there is more “weight” applied to the AVC component.
There was nothing in the debate of CVC vs AVC that adjusting the CVC and AVC charges could reduce the cost of building and running the network. At all, and anyone who tries to suggest that there was should get better informed on what has been said.
*This was the premise of Simon Hackett’s arguement, he further went on to state that if you want to reduce this problem and stay with a higher number POIs you need to rebalance the CVC and AVC charges such that there is more “weight” applied to the AVC component.*
if you actually understood Hackett’s example, you’ll realise that he’s being disingenuous in suggesting that it’s possible to rebalance the CVC and AVC charges in the way he suggests without reducing NBN Co.’s total revenue. the inconsistency or contradiction in his assertions lies in the fact that he talks about the need to provision for 100mbps subscribers in one breath and then proceeds to analyse the revenue impacts solely in terms of 12/1 plans.
*There was nothing in the debate of CVC vs AVC that adjusting the CVC and AVC charges could reduce the cost of building and running the network. At all, and anyone who tries to suggest that there was should get better informed on what has been said.*
mate, i was just trying to make sense of the quoted testimony. this is what happens when you’re dealing with NBN Co. bullshit and spin. the fundamental problem with the NBN is that it costs too much. the current pricing model shunts all the costs into the CVC component, so Ms Gillard can stand in front of the cameras and pronounce that “entry-level 12/1 plans will only cost $24”. reducing the number of POIs from 121 to 14 will not reduce the size of NBN Co.’s balance sheet. to the contrary, by incorporating more backhaul into the NBN, CVC charges will have to rise. and most of the industry complaints so far have centred on CVC charges.
you should keep better informed.
if you actually understood Hackett’s example, you’ll realise that he’s being disingenuous in suggesting that it’s possible to rebalance the CVC and AVC charges in the way he suggests without reducing NBN Co.’s total revenue. the inconsistency or contradiction in his assertions lies in the fact that he talks about the need to provision for 100mbps subscribers in one breath and then proceeds to analyse the revenue impacts solely in terms of 12/1 plans.
I understand Hackett’s example perfectly. I don’t agree that it is the right way to deal with the CVC charges, but at core it is a workable option, sure he may have been a little iffy in his modeling, but the concept of moving costs from one bucket to another is sound. It won’t reduce the overall cost of the NBN, it won’t reduce the amount that ISPs have to pay, but it will increase the amount they can deliver given the same wholesale payment input by reducing the amount of redundant charges. This will result in them being able to provide a better selection of plans for users.
mate, i was just trying to make sense of the quoted testimony. this is what happens when you’re dealing with NBN Co. bullshit and spin. the fundamental problem with the NBN is that it costs too much. the current pricing model shunts all the costs into the CVC component, so Ms Gillard can stand in front of the cameras and pronounce that “entry-level 12/1 plans will only cost $24″. reducing the number of POIs from 121 to 14 will not reduce the size of NBN Co.’s balance sheet. to the contrary, by incorporating more backhaul into the NBN, CVC charges will have to rise. and most of the industry complaints so far have centred on CVC charges.
No you weren’t. You were putting words in Quigley’s mouth. You were trying to say the less POIs will reduce the cost of the NBN, which Quigley didn’t say did he?
you should keep better informed
You should stop assuming that because I have come to a different conclusion to you I haven’t been paying attention.
*but at core it is a workable option*
workable, how? please explain. i would like to see your detailed calculations (as opposed to more back-tracking gibberish).
*but the concept of moving costs from one bucket to another is sound.*
again, since you make the claim… your detailed calculations, thanks. i’d like to see you out-do the Internode MD.
*It won’t reduce the overall cost of the NBN, it won’t reduce the amount that ISPs have to pay, but it will increase the amount they can deliver given the same wholesale payment input by reducing the amount of redundant charges. This will result in them being able to provide a better selection of plans for users.*
save me the gibberish and show me your calculations.
You have two buckets of water (VC costs), you’re trying to get X water to a location (the amount you need to pay per head of population on average in order to repay the debt NBN Co has), you have Y water in one bucket (Projected Revenue for AVC costs), you have Z water in another (Projected Revenue from CVC costs) such that X = Y + Z (Total Revenue from VC costs). You increase the amount of water in Y (You increase the cost of AVC) and decrease the amount of water in Z (You decrease the cost of CVC) such that X remains the same. This isn’t actually complicated maths Tosh. The CVC and AVC prices are completely arbitrary you realise?
That gibberish is the freaking core of the entire argument. The point that Simon Hackett made in clear and simple terms is that if you want to deliver 100Mbps connections you need to provide at least 200Mbps worth of CVC to the set of customers at a given POI in order to ensure that they get burst speeds of 100Mbps. That means that if you have only a small set of customers at a given POI that is less than the amount you would need to match your contention ratio for CVC costs your per customer cost goes up, a cost you will likely directly pass onto the customers.
However, if you reduce the CVC charges and increase the AVC charges, the amount it costs to deliver the minimum CVC in order to provide the correct burst speeds goes down, which in turn reduces the extra amount you will need to pay when you have customers less the ideal amount needed to match your desired contention ratio for CVC, meaning that extra cost you will likely directly pass onto customers is reduced.
Once you understand this, the problem just becomes finding the correct “balance” between AVC and CVC charges such that NBN Co will get enough revenue. And that, goes straight back to the X = Y + Z problem I just outlined above doesn’t it?
PIR %CVC total AVC CVC CR $/Mbit
12 9.1 26.4 24 2.4 0.01 20
25 27.0 37 27 10 0.02 20
50 46.9 64 34 30 0.03 20
100 67.8 118 38 80 0.04 20
250 78.1 320 70 250 0.05 20
500 85.7 700 100 600 0.06 20
1000 90.3 1,550 150 1,400 0.07 20
(i) start off with the 12/1 plan: assuming a contention ratio of 100:1, CVC charges of $2.40 are small relative to AVC charges of $24. hence, it’s trivial to bump up the 12/1 port cost by a dollar or two and reduce the CVC unit cost from $20/Mbit to $1/Mbit without causing any major drama.
(ii) now, cast your eyes up the speed tiers: CVC charges increasingly dwarf port costs. so, for the higher speed tiers, it’s not so easy to play the game of shifting costs from the “CVC bucket” into the “AVC bucket”.
(iii) you can factor in a more moderate fall in contention ratios across the higher speed tiers – it doesn’t change the underlying conclusions.
you still think it’s possible for NBN Co. to reduce the unit cost of CVC from $20/Mbit to $1/Mbit without dramatically inflating port costs for the higher speed tiers?
Now you’re shifting the goal posts. Since when did either of us specify the condition of not dramatically increasing the AVC costs since that is now the only, very thinly vailed arguement. I am well aware the in order to keep the access costs per user for
Now you’re shifting the goal posts. Since when did either of us specify the condition of not dramatically increasing the AVC costs per user since that is now the only, very thinly vailed argument. I am well aware the in order to keep the access costs per user for the different speed tiers will likely result in a dramatic inflation of the costs of the higher tier speeds, because I specified the goal of mandating the the overall cost the same.
I have done the Math, and using a different set of assumptions to you (mainly because I think an 100:1 contention ratio should never, ever, be used by an ISP in industry). Uniform contention of 20:1. Under the current model ($20/Mbit CVC) I get the following: (apologises for the difficult to read nature of it)
PIR (Mbit, Down) PIR (Mbit, Up) AVC ($/m Per User) AVC Total ($/m) CVC (Mbit) CVC ($/m) Total ($/m) Per User Total ($/m) % CVC
12 1 24.00 720.00 13 260.00 980.00 32.67 26.53%
25 5 27.00 810.00 30 600.00 1,410.00 47.00 42.55%
25 10 30.00 900.00 35 700.00 1,600.00 53.34 43.75%
50 20 34.00 1,020.00 70 1,400.00 2,420.00 80.67 57.85%
100 40 38.00 1,140.00 140 2,800.00 3,940.00 131.34 71.07%
250 100 70.00 2,100.00 350 7,000.00 9,100.00 303.34 76.92%
500 200 100.00 3,000.00 700 14,000.00 17,000.00 566.67 82.35%
1000 400 150.00 4,500.00 1400 28,000.00 32,500.00 1,083.34 86.15%
Now, droping that down to a reasonable amount of say $5/Mbit for CVC:
PIR (Mbit, Down) PIR (Mbit, Up) AVC ($/m Per User) AVC Total ($/m) CVC (Mbit) CVC ($/m) Total ($/m) Per User Total ($/m) % CVC Var
12 1 30.00 900.00 13 65.00 965.00 32.17 6.74% -0.50
25 5 42.00 1,260.00 30 150.00 1,410.00 47.00 10.64% –
25 10 48.00 1,440.00 35 175.00 1,615.00 53.84 10.84% 0.50
50 20 69.00 2,070.00 70 350.00 2,420.00 80.67 14.46% –
100 40 108.00 3,240.00 140 700.00 3,940.00 131.34 17.77% –
250 100 245.00 7,350.00 350 1,750.00 9,100.00 303.34 19.23% –
500 200 450.00 13,500.00 700 3,500.00 17,000.00 566.67 20.59% –
1000 400 850.00 25,500.00 1400 7,000.00 32,500.00 1,083.34 21.54% –
Notice that although the overall cost per user remains the same except in the 25/5 case, the AVC costs have jumped considerably, for example to get the 1000/400Mbps connection you now need to pay $850 in AVC. The problem is considerably more complicated in practice because there will be different contention ratios for different plans, etc.
So, my goal was to reduce the impact of redundant CVC. Sticking to the the 1000/400 example, if you provided this service at 20:1 contention, right now, if you only have 2 subscribers via the POI, you will need to likely provide 2000Mbps of CVC. This is very likely because it is unlikely many users could afford $1100 a month for Broadband. In the current, $20/Mbit environment, to do this would cost you, the RSP, an extra $37,800/m to provide these services to those customers. Meaning you likely just won’t bother, or if you do, you’ll pass those costs directly onto the customer.
But sure, you move the goal posts, because obviously Ms Gillard being able to cite AVC costs to show just how cheap the NBN will be to run is more important than actually ensuring that RSPs don’t end up overcharging customers because they are lumped with CVC charges of non-trivial amounts if they don’t meet a minimum number of customers per POI.
I think I remember saying at some point that the best way to reduce the costs of NBN services to customers and wholesalers was actually to sink money into it rather than to expect a return. Fluffing around with the CVC and AVC costs isn’t going to change this fact.
*brain hurts* I almost failed math in first year uni …
i’m sorry, immensely complicating the calculations achieves what?*
look, you have to realise that, given a choice between dealing with a non-structurally-separated Telstra FTTN/copper network and dealing with a wholesale-only, government-owned FTTP network, Telstra’s competitors, whose business models revolve around access to infrastructure they don’t own, will generally pick the latter any day.
all these ISPs care about is preserving and growing their gross margins. achieving this objective is not necessarily 100% aligned with consumer interests of cheap, affordable broadband. while high CVC charges may kill the higher gross margins earned from flogging high quota plans, Telstra’s competitors are clearly tantalised at the prospect of growing their gross margins [in a different dimension] by capturing more market share in the regional areas (which is still heavily-dominated by Telstra).
Vodaphone’s expressed interest in participating as an RSP on the new fibre network is not a vote of confidence in the NBN (like the Minister likes to spin it). it’s merely a competitive response or commercial recognition of the fact that the superior competitive advantage of “product-bundling strategies” will be further entrenched under the new fibre regime thanks to NBNco’s wholesale product configuration and pricing model.**
the fundamental problem with Hackett’s numerical exercise is that by solely confining the calculations to the entry-level plans, it (intentionally or unintentionally) gives the misleading impression that the “exorbitant cost” issue of building 93% FTTP can be dealt with simply by shifting a few dollars “from one bucket to another bucket”. this is clearly not the case.***
there’s nothing more to it than that. stop arguing for the sake of arguing.
* btw, on a trivial note, your numbers do not imply a uniform contention ratio of 20:1. instead, they run as high as 30:1 for the 12/1 plan.
** the whole NBN project is largely devoid of any economic rationale. instead, it’s solely a “reverse engineering” exercise to satisfy two arbitrary “political edicts”, viz. build FTTP and 7% ROI. in the course of achieving these two questionable goals, all other important considerations have been cast aside or thrown under the wheel. add to that the “organisational arrogance” that comes naturally from running a freshly-minted, government monopoly… and you have the unfolding disaster that is the “NBN”.
*** to be fair, this “small issue” aside – by speaking out instead of remaining silent like other ISPs, Internode have brought heaps more clarity to the marketplace on the current flaws of the NBN business model. (of course, like everyone else, everything they do is motivated by self-interest.****)
**** while, on paper, building a wholesale-only, fibre platform opens up the regional market for Telstra’s competitors, i think it’s quite obvious that ISPs like Internode are disturbed (and envious) at the (unavoidably) high level of co-operation and “joint-venturing” between NBNco and Telstra and the competitive advantage this gives the #1 telco in terms of market intelligence/planning and defending its current market share. Internode’s probably hoping for more “buy-in” in this whole complicated process of “copper-to-fibre” migration.
* btw, on a trivial note, your numbers do not imply a uniform contention ratio of 20:1. instead, they run as high as 30:1 for the 12/1 plan.
Apologises. I actually did do the contention ratio at 30:1 it seems looking back at the spreadsheet constants. Just replace the two occurrences of 20:1 with 30:1.
i’m sorry, immensely complicating the calculations achieves what?*
I’m sorry what? How are my calculations complicated. I just adjusted the AVC and CVC charges until I got the same wholesale cost per user with a reduced CVC. To actually adjust the AVC and CVC charges accruately would require more in-depth analysis. And for the record, you did ask me for calculations on the issue, I gave them to you, and now you want me to make the simpler for you?
look, you have to realise that, given a choice between dealing with a non-structurally-separated Telstra FTTN/copper network and dealing with a wholesale-only, government-owned FTTP network, Telstra’s competitors, whose business models revolve around access to infrastructure they don’t own, will generally pick the latter any day.
So that’s a point in favour of a government-owned FTTP network is it? Or are you trying to imply that their business models are broken? I don’t quite understand what the point of this particular statement was.
all these ISPs care about is preserving and growing their gross margins. achieving this objective is not necessarily 100% aligned with consumer interests of cheap, affordable broadband. while high CVC charges may kill the higher gross margins earned from flogging high quota plans, Telstra’s competitors are clearly tantalised at the prospect of growing their gross margins [in a different dimension] by capturing more market share in the regional areas (which is still heavily-dominated by Telstra).
Okay, and reducing the CVC charges reduces the amount of wholesale cost they have to pay per customer, which in turn allows them to reduce charges while maintaining their margins. Which is the reason why an adjustment to the ratio of CVC to AVC may be in order to ensure that maintaining this business model does allow them to achieve the objective with consumer interests.
Vodaphone’s expressed interest in participating as an RSP on the new fibre network is not a vote of confidence in the NBN (like the Minister likes to spin it). it’s merely a competitive response or commercial recognition of the fact that the superior competitive advantage of “product-bundling strategies” will be further entrenched under the new fibre regime thanks to NBNco’s wholesale product configuration and pricing model.**
So what? Me pointing out how the adjusting the CVC and AVC prices works to you is not an endorsement for the NBN either, I did mention at the bottom that I think the easiest way to fix the CVC cost problem is to directly sink money into the project.
the fundamental problem with Hackett’s numerical exercise is that by solely confining the calculations to the entry-level plans, it (intentionally or unintentionally) gives the misleading impression that the “exorbitant cost” issue of building 93% FTTP can be dealt with simply by shifting a few dollars “from one bucket to another bucket”. this is clearly not the case.***
I just explained to you above how it still works in the higher tiers with figures to back it up, and you’re still trying to tell me that you just can’t adjust the CVC and AVC charges…. because why? You just can’t?
there’s nothing more to it than that. stop arguing for the sake of arguing.
If that is the case why do you keep making stuff up and shifting your goal posts? First you say that the adjusting the costs won’t reduce the cost of the NBN, I point out that wasn’t the case and it was about reducing the amount of redudant CVC charges when there are a low amount of users on a POI, and then you try and tell me that we need to do that without adjust AVC charges?
*How are my calculations complicated.*
they don’t illustrate anything new.
*And for the record, you did ask me for calculations on the issue, I gave them to you*
you mean you replicated my calculations after you failed to produce the numbers i asked for.
*and now you want me to make the simpler for you?*
don’t make your “modelling” any more complicated that it has to be – unless it’s merely a “face-saving” attempt.
*I don’t quite understand what the point of this particular statement was.*
the first 4 words of that sentence sums up your participation in this discussion.
*Okay, and reducing the CVC charges reduces the amount of wholesale cost they have to pay per customer, which in turn allows them to reduce charges while maintaining their margins.*
oh dear…. for the umpteenth time.. this dollar-shifting game seems facile and trivial when you’re dealing with 12/1 plans with 100:1 contention ratios because there isn’t much CVC provisioning in the first place! but, once you start to cast your eyes across the higher speed tiers, this “bucket-swapping” approach clearly doesn’t work.
in other words, the access charges for the NBN have two components: a fixed tariff (AVC) and a variable tariff (CVC). in the case of the 12/1 plan with 100:1 contention ratio, the variable tariff is so trivial that it can be absorbed within the fixed tariff without dramatically inflating port charges. however, once you start migrating up the higher speed tiers, the variable tariff becomes so substantial that it’s no longer possible to simply absorb the CVC charges within the fixed tariff without fucking up your port charges. HACKETT’s SUGGESTION DOESN’T WORK.
ffs…
*I just explained to you above how it still works in the higher tiers with figures to back it up, and you’re still trying to tell me that you just can’t adjust the CVC and AVC charges…. because why? You just can’t?*
do you even understand the concept and underlying rationale of a “two-part tariff”? obviously not.
if you cram or collapse CVC charges into fixed port fees, what flexibility does that allow you in terms of offering different throughput allowances and contention ratios to your customers?
*If that is the case why do you keep making stuff up and shifting your goal posts?*
you’re obviously a highly-intelligent person – stop making a fool of yourself in this thread.
How pleasing to see the wholesaler and RSP discusing how to make the model function. With goodwill on both sides a win-win outcome would be assured.
Are you the guy who nods knowingly in the background when Gillard, Swan, Conroy etc are being interviewed for the news?
Gee Visionary, seems your own team turn on you easily eh?
Even when you are only being facetious…
One might say they are most “typically visionless”…!
The vast majority of POIs in the 121 POI setup are suburbs in major cities. There are only a small number of POIs in regional locations. So the extra cost in connecting a 14 POI network is relatively small. After all the majority of the cost in the NBN is all the house to house cabling.
The REAL problem here is that NBNco is charging for the size of the pipe and NOT the actual number of bits used. Net result is that if you are a small RSP and you have fewer users per POI, then each “pipe” you have to pay for actually costs your more per GB of actual data. This is because when you have only a few users you cannot efficiently use the capacity of your pipe.
The other problem is the sheer scale of the operation and all the setup costs in all of the locations and all of the independent deals you have to do to get backhaul to all of the POIs. If you’re a small ISP/RSP with just a few thousand users then you’re in a world of pain paying for all these costs multiplied a hundred times. NBNco seems to be working with past precedents – that ISPs start off small and local and then expand. Yes, this is true, and in theory it can be made to happen on the 121 POI network. But this ignores totally the possibility of RSPs catering to niche markets that are geographically diffuse.
Mr. Hackett is quite right – it will put pressure on start ups (and by implication a lot of present smaller ISPs) to pay extra to go through a wholesaler.
Why then cant NBNco can’t either simply charge per GB directly (rather than provide a pipe that you then have to find enough users to fill) After all, the pipe is a virtual construct anyhow. This would certainly help.
Nor do I get why NBNco can’t do a reverse volume discount – meaning charging less for CVC in situations where a RSP has relatively few users.
*Why then cant NBNco can’t either simply charge per GB directly (rather than provide a pipe that you then have to find enough users to fill) After all, the pipe is a virtual construct anyhow. This would certainly help.*
on a general level, as the network owner operator, NBN Co. has to engage in forward network planning to ensure the multitude of users of the network have sufficient capacity allocated to them in order for things to run smoothly. this is best facilitated by access seekers signalling in advance how much capacity provisioning they require on an individual basis.
it’s easiest and most convenient to auction off slices of the aggregate bandwidth capacities of individual pipes by selling “ex-ante capacity” on a “time-dependent” basis, rather than charging for “ex-post throughput” on a “temporally-unbundled” basis. it makes it simpler for them to price congestion and manage the network.
*Nor do I get why NBNco can’t do a reverse volume discount – meaning charging less for CVC in situations where a RSP has relatively few users.*
by implementing a “reverse volume discount”, you’re basically suggesting that NBN Co. should “oversubscribe” CVC capacity notionally allocated to particular RSPs with “relatively few users”. because these RSPs do not have a sufficiently large user base to amortise the CVC overhead, NBN Co. can “oversell” the CVC capacity and pass on the “extra” revenue (or “savings”) to these RSPs through “discounts”.
hence, you’re basically suggesting that NBN Co. should act as a wholesale aggregator.
Yes, of course its more convenient for NBNco to sell capacity rather than actual data. However its not a huge ask for NBNco to do just that and in practice it will initially heavily over-provision anyhow meaning it has plenty of time to spot trends and adjust its equipment as necessary.
No, what I’m suggesting with a reverse volume discount is that NBNco discount somewhat its CVC charges for RSPs in situations where the RSP hasn’t yet built up a critical mass of users on a given POI (measurable in terms of ordered AVCs).
“hence, you’re basically suggesting that NBN Co. should act as a wholesale aggregator.”
To some extent yes. Its an idea. I see the logic in wholesale aggregators. The problem is are we going to get aggregators that simply hide the details of the backhaul and CVC and allow the RSP to purchase the port from NBNco and carry on as usual, or are we going to get wholesalers that effectively resell ports?
In any case, there are some essential costs with dealing with lots of POIs that just simply aren’t really necessary. We’ll see I guess..
*No, what I’m suggesting with a reverse volume discount is that NBNco discount somewhat its CVC charges for RSPs in situations where the RSP hasn’t yet built up a critical mass of users on a given POI (measurable in terms of ordered AVCs).*
there’re two different dimensions to this “small ISP” problem that shouldn’t be confused:
(i) during the “roll-out phase” of the NBN, everyone will be a “small ISP” on the partially-completed fibre network. this is the so-called “slow walk along the valley of death” problem that Internode has christened.
(ii) this is separate from the viability issue of genuine “small ISPs” following the complete migration of all customers from copper to fibre.
there’s obviously some merit in NBNco temporarily “discounting” CVC charges during the ramp-up phase of the fibre network as ISPs gradually fill their POIs with their migrated copper customers. it’s hard to imagine this won’t have some effect on NBNco’s revenue projections. however, given that CVC revenue is relatively small during the early days of the network, some form of temporary “discounting” might be “workable”.
this is quite different from giving permanent “discounts” to small ISPs following the complete migration to fibre. in the absence of these “special deals”, these small ISPs would be accessing the NBN via a wholesale aggregator. so, by dealing directly with these small ISPs and assigning them the same quanta of shared CVC capacity that these ISPs would otherwise get from a wholesale aggregator, NBNco’s effectively acting as a wholesale aggregator. remember, all these “pricing rearrangements” have to be revenue-neutral to NBNco.
NBNco will never do this as it forces them to adopt more roles, expand its functions and take on additional types of risk that fall outside its original mandate.
*The problem is are we going to get aggregators that simply hide the details of the backhaul and CVC and allow the RSP to purchase the port from NBNco and carry on as usual, or are we going to get wholesalers that effectively resell ports?*
i don’t see a difference. of course, the wholesale aggregators will extract a margin on their services and these small ISPs will be hostage to the capacity sharing and pricing models imposed by the large aggregators.
another way of looking at the problem is this:
the issue with the $20/Mbps CVC charge is not so much that it starts at $20/Mbps – it’s the fact that it starts at $20/Mbps and stays there. in other words, the cost of data is linear, as opposed to logarithmic – there’s no volume discounting.
importantly, from the perspective of NBNco, the “barriers to entry” for small ISPs that result from a high starting point for CVC charges is actually potentially a good thing. as the network owner operator, you prefer to deal with the big telcos such as Telstra, Optus, iiNet, Internode, etc. you don’t want to deal directly with zillions of small little ISPs because:
(i) it results in too much operational and administrative complexity;
(ii) dealing directly with many small, undercapitalised ISPs is operationally and financially riskier than dealing with large, well-capitalised telcos.
this is why it actually makes sense to offer special or concessionary pricing terms to large telcos or wholesale aggregators; because, by acting as intermediaries between NBNco and the rest of the market, they actually provide a valuable service by removing an unnecessary layer of financial and operational risks from NBNco. it hugely simplifies NBNco’s network administration and minimises the role of the government monopoly in the marketplace in terms of active “risk-taking”.
real world markets are characterised by complexity, heterogeneity and diversity. the whole NBN model of a “single platform”, “single market” and “single price” and the massive economic inefficiencies that result is quite frankly socialist madness of the first order.
A further example of the stupidity of the ACCC’s decision on POIs.
The original NBN costings and construction timeframe envisaged a simple and robust regional network of leased or built longhaul fibre with multiple paths back to two geographically separated redundant POIs in each state capital.
No single break would cause an outage to regional users. Small and big RSPs could serve customers anywhere with cheap capital city connections from NBNCo’s POI to their own backhaul. Telcos with longhaul fibre investments would gain a new customer in NBNCo, to add to their big VPN users in government and big business.
But no – telco and coalition lobbyists got the ACCC to arbitrarily require more than a hundred extra POIs, exacrebating network complexity and adding the cost burden of commercial longhaul fibre on small RSPs, eliminating them as serious competitors to Telstra, Optus and Soul, etc.
Hilariously, the argument went that this would benefit small operators, and protect the investment of longhaul fibre owners, who in reality would get NBNCo as a customer under either POI model.
Whatever the eventual ratio of AVC and CVC costings, the additional cost of building and operating 108 extra NBNCo facilities nationwide will be added to the cost of doing business for all telcos, and it will especially hurt RSPs that do not own longhaul fibre to their target markets.
The ACCC is to review the POI count with NBNCo in June. Let us hope that sanity prevails and NBNCo’s cheaper, more competitive and technically-superior network design gets its guernsey back.
*Telcos with longhaul fibre investments would gain a new customer in NBNCo, to add to their big VPN users in government and big business.*
in general terms, if you’re moving from a “5000 exchange” model to a “121 POIs” model, some backhaul is bound to be stranded. this is especially the case if you further compress the number of POIs from 121 to 14. so, the value of existing backhaul is highly-dependent on the number of POIs and where these POIs are located.
*But no – telco and coalition lobbyists got the ACCC to arbitrarily require more than a hundred extra POIs, exacrebating network complexity and adding the cost burden of commercial longhaul fibre on small RSPs, eliminating them as serious competitors to Telstra, Optus and Soul, etc.*
RSPs will have to foot the cost of “longhaul fibre” regardless of whether this “backhaul” is purchased as a package from “14 POIs NBN” (i.e. even higher CVC charges) or leased separately from third parties. however, it’s true that the cost of installing and maintaining equipment in 121 POIs as opposed to 14 POIs increases their overhead (assuming they want to serve all POIs).
basically, if you view NBN Co. as some benevolent Father Christmas, with unlimited taxpayer bankroll, magically dispensing “cheap fibre” to all and sundry, then you’ll prefer NBN Co. to provide as much backhaul as possible.
however, realists (i.e the “telco and coalition lobbyists”) recognise that NBN Co. will have a massive balance sheet (or capital structure) to service. the release of NBN Co.’s corporate plan has already confirmed that they will be charging monopoly rents (i.e. $20/Mbps CVC) on the infrastructure they control (i.e. fibre CAN up to 121 POIs) in order to make their magical 7% ROI.
the fear was always that if NBN Co. was allowed to reach further backwards into more intra-state backhaul (up to the 14 POIs in capital cities, for instance), it would merely hand this 900-lb, government-owned gorilla even more infrastructure over which it could extract its monopolistic rent. that would be highly-dangerous and most undesirable.
*Hilariously, the argument went that this would benefit small operators*
the argument here was that you could have regional ISPs hooking up the local town council with the local public library via a nearby POI. this would not be possible if all POIs were located in the capital cities. under the 14 POIs model, connecting the local council with the local library via a capital city POI would possibly result in greater latency than connecting a Sydney server to a NZ server.
*the additional cost of building and operating 108 extra NBNCo facilities nationwide will be added to the cost of doing business for all telcos*
all these 121 POIs are basically refurbished Telstra exchanges. for the industry as a whole, going from 5000 exchanges to 121 POIs should result in some savings. also, you have to realise all the complaints so far from industry have centred on CVC charges (which NBN Co. has made plain clear isn’t a backhaul charge) and not the NNI charges (associated with each POI). more importantly, even if you reduce the number of POIs, the financial burden of CVC charges on RSPs in absolute dollars terms will still be the same. what NBN Co. loses in the swings will have to be offset in the roundabouts in order to ensure income-neutrality (i.e if RSPs purchase half as many CVCs under 14 POIs, NBN Co. will have to double CVC charges to $40/Mbps to maintain revenue forecasts).
*The ACCC is to review the POI count with NBNCo in June. Let us hope that sanity prevails and NBNCo’s cheaper, more competitive and technically-superior network design gets its guernsey back.*
there’s nothing “competitive” about having a giant, government-owned monopoly reaching backwards and absorbing all the intra-state backhaul.
I have never, ever heard of a wholesale network employing CVC charges. That has got to be the stupidest pricing plan of all time in the internet industry. Thanks to CVC the vast majority of the fiber network’s capacity will go unused. AVC alone should be enough. If you want to lower prices for the lowest end, then raise prices on the higher end plans. How hard is that? The people who want extremely high bandwidth will no doubt pay premium prices for it.
CVC charges are also insanely high. In a couple years international transit will be less expensive than CVC. What kind of an idiot pricing plan is that?
“CVC charges are also insanely high. In a couple years international transit will be less expensive than CVC. What kind of an idiot pricing plan is that?”
It’s a plan to try and recoup some revenue from the most expensive FTTH rollout in the World, paid out of the public purse then given back to private industry to resell at a hefty markup to the sucker taxpayers that paid for it in the first place.
Uh… it’s not the most expensive rollout in the world, and since all wholesale revenue will go to NBNCo the cost is easily manageable. You’ve demonstrated yourself to be a hypocrite time and time again on this forum so calling others “idiots” is pathetic.
alain said… “it’s a plan to try and recoup some revenue from the most expensive FTTH rollout in the World, paid out of the public purse then given back to private industry to resell at a hefty markup to the sucker taxpayers that paid for it in the first place”.
“GIVEN” back to private enterprise…? Please…
Intentional, disgraceful FUD if I have ever seen it…’given”! Whether you are referring to RSP’s “paying” to access the NBN, or the network itself being privatised/”sold”… that is simply typical alain lies, to which we are all accustomed!
How about… paid out of the public purse (because private enterprise refuses) then “wholesaled” to the financial advantage of the taxpayer, to private industry to resell, as our businesses do (and yes they will make a profit – please don’t scream free market then “contradictorily… surprise, surprise”, have the audacity to whinge because companies make a profit…FFS) to customers “exactly as they do now”. But with the NBN, they will be reselling to a lot of people whom would otherwise never have decent, affordable comms. Plus the best of all for the fiscally anal, it will over time, “pay for itself”, costing the taxpayer …$0
Please ignore the above facts and talk.. 7% ROI, LOL, WGAF? Paid back in full = cost $0 (which you blindly ignore)!
But not only will it cost $0, but add the economic benefits (again blindly ignored by you…) and then the sell of price (surprise surprise ignored and/or refuted by you…)… sigh!
The we have YOUR plan –
Gift (yes gift this time is most apt) private enterprise $B’s from the public purse, to build and own our network and sell it back to us (possibly at gouging, pre-NBN prices) and the sucker taxpayer will NEVER receive even 1c in return…
Yes brilliant…NOT!
they don’t illustrate anything new.
*clap*…*clap*…*clap*. Oh good that’s still working. – GLaDOS.
So nothing new somehow equals complexity now does it? When all I was trying to do was explain the logic behind Hacketts position. Which is all you asked of me. And have continued to ask of me.
you mean you replicated my calculations after you failed to produce the numbers i asked for.
No I didn’t replicate your calculations, I created my own calculations, based upon yours, to show the concept of inflating the AVC costs to compensate for the drop in CVC costs. So now adding to is somehow considered replication now is it?
don’t make your “modelling” any more complicated that it has to be – unless it’s merely a “face-saving” attempt.
I didn’t. I actually made it quite simple. I removed some of the variables from the equation (e.g. variance of contention) in order to quickly show how inflating the AVC costs can allow for deflation of the CVC costs. Again, providing you exactly what you asked me for.
Let’s actually look at what you said:
workable, how? please explain. i would like to see your detailed calculations (as opposed to more back-tracking gibberish).
Which I answered with a simple analogy of moving water between buckets. Followed by:
you still think it’s possible for NBN Co. to reduce the unit cost of CVC from $20/Mbit to $1/Mbit without dramatically inflating port costs for the higher speed tiers?
To which I replied a flat out NO. Followed by a series of calculations to show you this inflation does occur an explanation as to why this may be desirable over the current model (reducing the amount of redundant changes ISPs have to pay if they have a POI that they not servicing enough to actually get the benefits of reduced CVC charges due to provided a contented service). Followed by:
i’m sorry, immensely complicating the calculations achieves what?*
At which point… well. Let’s see it again shall we?
If that is the case why do you keep making stuff up and shifting your goal posts? First you say that the adjusting the costs won’t reduce the cost of the NBN, I point out that wasn’t the case and it was about reducing the amount of redudant CVC charges when there are a low amount of users on a POI, and then you try and tell me that we need to do that without adjusting AVC charges?
I have carefully read everything you had said and so far you have said that “bucket swapping doesn’t work because it’s no longer a trivial problem when you start looking at the higher speed tiers”. Which in itself doesn’t tell me why adjusting the CVC and AVC costs can’t be done. You are making a logical leap here which you are either failing to explain, or relies on some strange definition of “workable solution” that implies triviality.
oh dear…. for the umpteenth time.. this dollar-shifting game seems facile and trivial when you’re dealing with 12/1 plans with 100:1 contention ratios because there isn’t much CVC provisioning in the first place! but, once you start to cast your eyes across the higher speed tiers, this “bucket-swapping” approach clearly doesn’t work.
in other words, the access charges for the NBN have two components: a fixed tariff (AVC) and a variable tariff (CVC). in the case of the 12/1 plan with 100:1 contention ratio, the variable tariff is so trivial that it can be absorbed within the fixed tariff without dramatically inflating port charges. however, once you start migrating up the higher speed tiers, the variable tariff becomes so substantial that it’s no longer possible to simply absorb the CVC charges within the fixed tariff without fucking up your port charges. HACKETT’s SUGGESTION DOESN’T WORK.
Did I say it was a trivial problem? No. I showed an example of what he means and purposely simplified the constraints of the problem in order to illustrate it. That is all.
What part of the following do you not understand: The current CVC and AVC prices as outlined by the NBN Co Business case are arbitrary. There is no in depth analysis of the rational behind them and the consequences of using these charges. It is therefore entirely possible that in provisioning for these prices that they made a mistake that could be better services by an adjustment of the CVC and AVC prices. To modify the prices is obviously not a trivial matter if you want to ensure that they will allow the company to continue to meet it’s stated return of 7%.
do you even understand the concept and underlying rationale of a “two-part tariff”? obviously not.
if you cram or collapse CVC charges into fixed port fees, what flexibility does that allow you in terms of offering different throughput allowances and contention ratios to your customers?
You are taking the argument to the extreme here. Further than Simon Hackett suggested it should be taken, further than I would suggest it be taken based upon my analysis and I decided there wasn’t a better pricing model to explore. If you reduce CVC to triviality, or remove it completely, you do remove the flexibility of different contention ratios in order to provide a range of products for different users. This is obvious.
you’re obviously a highly-intelligent person – stop making a fool of yourself in this thread.
Thank you for the backhanded compliment.
Sorry that was in reply to this comment.
*So nothing new somehow equals complexity now does it?*
i’m glad you finally agree your “face-saving calculations” don’t contribute anything new.
*When all I was trying to do was explain the logic behind Hacketts position. Which is all you asked of me. And have continued to ask of me.*
you claimed that Hackett’s solution is a “workable option” and that the “concept of moving costs from one bucket to another is sound”. i’ve already shown that it’s neither a “workable option”, nor is the “bucket-swapping” pricing game “sound”.
this essentially boils down to the fact that the “buckets” are NOT equivalent and perform different functions in the pricing equation. one is a “fixed tariff” while the other is a “variable tariff”. in the special case of a 12/1 plan with a 100:1 contention ratio, the level of CVC provisioning is so trivial that you can absorb the “variable tariff” into the “fixed tariff” without causing any drama.
however, once you cast your eyes over the entire spectrum of speed tiers, you quickly realise that the CVC charges become increasingly substantial and can no longer be disguised within the “fixed tariff” component of the pricing equation without fucking up your port fees.
hence, Hackett’s solution DOES NOT WORK.
*in order to quickly show how inflating the AVC costs can allow for deflation of the CVC costs. Again, providing you exactly what you asked me for.*
i didn’t ask you for calculations to verify the obvious fact that reducing NBNco’s revenue in the form of CVC charges will necessitate a higher take in port charges. this is precisely why your calculations were unnecessary.
*you have said that “bucket swapping doesn’t work because it’s no longer a trivial problem when you start looking at the higher speed tiers”. Which in itself doesn’t tell me why adjusting the CVC and AVC costs can’t be done.*
for the umpteenth time: “bucket-swapping” or shifting variable “CVC charges” into fixed “AVC fees” doesn’t work because it fucks up the pricing mechanism or underlying purpose of a “two-part tariff”. it only “works” if you pretend the higher speed tiers don’t exist.
*You are making a logical leap here which you are either failing to explain, or relies on some strange definition of “workable solution” that implies triviality.*
you mean you either fail to understand simple logical arguments or acknowledge the existence of higher speed tiers. or maybe you’re just being a total d!ckhead.
*The current CVC and AVC prices as outlined by the NBN Co Business case are arbitrary.*
what’s arbitrary about the NBN is the decision to build 93% FTTP and to “only” make a return of 7%. these “political parameters” determine the size of NBNco’s balance sheet and the revenue streams the project needs to generate.
view in this context, the AVC and CVC charges proposed by NBNco are NOT arbitrary. they’ve been finely-calibrated to satisfy these “political parameters”.
*There is no in depth analysis of the rational behind them and the consequences of using these charges. It is therefore entirely possible that in provisioning for these prices that they made a mistake that could be better services by an adjustment of the CVC and AVC prices.*
there’s no rationale behind indiscriminately forcing (almost) everyone onto the fibre platform and the consequent impact on NBNco’s bloated capital structure. however, given these parameters, there’s definitely a solid rationale behind adopting a “two-part tariff” pricing mechanism (which is clearly something you’ve very little understanding of) in an attempt to recoup the costs.
*You are taking the argument to the extreme here. Further than Simon Hackett suggested it should be taken, further than I would suggest it be taken based upon my analysis and I decided there wasn’t a better pricing model to explore.*
i’m taking what to extreme? Mr Hackett has suggested that it’s feasible to reduce CVC charges from $20/Mbps to $1/Mbps. this is clearly and unambiguously NOT A WORKABLE SOLUTION. a solution is only “workable” if “works” for all speed tiers because the CVC charge applies to all speed tiers. stop being a d!ckhead.
*If you reduce CVC to triviality, or remove it completely, you do remove the flexibility of different contention ratios in order to provide a range of products for different users. This is obvious.*
yea, it’s OBVIOUS now that i’ve pointed it out to you – and you still maintain it’s a “workable solution”!
*Thank you for the backhanded compliment.*
i think i’ll have to retract my compliment after reading this latest rubbish post of yours. some people just don’t know how to make a graceful concession and accept that they’re wrong instead of engaging in disingenuous, dissembling forum lawyerism.
is it that difficult to say, “i was wrong”?
is it that difficult to say, “i was wrong”?
No, because I am actually agreeing with you and have been agreeing with you for quite a while. You just haven’t been listening. You are so focused on proving me wrong that you haven’t actually listened to what I have said and tried to repeat the same arguments.
The process of adjusting the CVC and AVC costs is not a trivial thing to do, and to do so will dramatically change the structures of plans that can be provided and runs the risk of not meeting the stated financial goals of the NBN. All things I have already said in this thread.
The reason that this debate is brought up being is that although the prices have been finally celebrated they were created on a false assumption, that false assumption was that there were would be less than twenty POIs. This is no longer the case and pricing has not changed to reflected this both because it isn’t a trivial thing to do. All Simon Hackett is trying to do is bring attention to this fact, and this is why NBN Co opening takes with him on this issue is a good thing.
there’s no rationale behind indiscriminately forcing (almost) everyone onto the fibre platform
Okay, this I don’t like about your debating style tosh: disagreement with the rational doesn’t mean there isn’t one. A rational that is wrong is still a rational and your goal should not be trying to show that there isn’t any rational, but that the current rational is wrong. For the most part you do this when discussing the NBN, almost by accident, in particular because the rational behind the NBN hasn’t been discussed in depth in a public forum.
however, given these parameters, there’s definitely a solid rationale behind adopting a “two-part tariff” pricing mechanism (which is clearly something you’ve very little understanding of) in an attempt to recoup the costs.
I didn’t say there wasn’t a rational, I said it wasn’t discussed in depth. There is a difference.
But since you overtly want me to say it, here ya go: the process of changing the CVC and AVC charges in order to insure flexible pricing and well as recoup of the debt for the NBN is so complicated that is probably not worth the effort, and as such we might want to consider a more creative solution to the problem. The solution I suggested would be to directly sink some of the government investment in the NBN rather than expect a return, which means you can cut down on the amount of income you need to make.
tsohP300, seriously…
You are simply embarrassing yourself… further…now!
Whilst we have had our moments of disagreement, I don’t like to see anyone, foolishly continuing a debate when they are obviously “totally lost”…(as you your self clearly admitted, already)!
Just quit while you’re are (again as you admitted) “totally lost” and before your rank rating of “low aura” (as it currently stands), lowers even more, if that is possible…
Don’t thank me for the friendly advice, it’s my pleasure!
Comments are closed.