‘Insane’ NBN pricing will kill small ISPs: Hackett


Update: Hackett has published a blog post with the slides from his talk, which are very instructive to his argument.

Internode managing director Simon Hackett today described the National Broadband Network’s pricing model as “insane” for small internet service providers, warning that none will survive their walk through the “valley of death” transition from the current copper network to the fibre future envisioned by the Federal Government.

According to Hackett, the NBN’s pricing model will only be feasible for ISPs with larger than 250,000 customers – which only five retail ISPs in Australia can boast – Telstra, Optus, TPG, iiNet and Internode itself, which will just reach the cut-off mark.

For ISPs with 250,000 customers or more, Hackett told the Communications Day Summit in Sydney this morning, it would cost as little as $27.28 per month plus GST, according to his calculations, to connect customers to the fibre infrastructure as it was rolled out.

However, the Internode chief said, the price rose dramatically as the number of customers fell – so that for a national ISP with only 10,000 customers spread out across Australia, the cost of connecting to the NBN would be as high as $106 per month per customer. “I just gave you insane,” he said. “At 10,000 customers, it’s insane to connect to this network, as a national provider. A 10,000 customer provider is insane to connect to this network.”

According to Hackett, there were several fundamental problems with NBN Co’s pricing model as laid out in its business plan published late in 2010. For example, he said, the decision by the Australian Competition and Consumer Commission to set the number of points of interconnect around Australia where ISPs could connect to the network at 120 nationally disadvantaged smaller ISPs – who would be forced to provide what NBN Co calls a Connectivity Virtual Circuit (CVC) connection to all of these locations in order to provide a national service.

Hackett has previously advocated a model where NBN Co would provide as little as 14 points of interconnect – which he said would serve smaller ISPs much better. A model with more points of interconnect would serve larger ISPs like Telstra and Optus, he said, which already had infrastructure around Australia.

Secondly, he said the cost of that CVC circuit was too high – it should come down dramatically. Instead, NBN Co should charge more per month for each individual customer connection.

If NBN Co’s pricing model didn’t change, Hackett argued, the only way which smaller ISPs would be able to afford to connect to the network would be through larger wholesale players like Optus and Telstra. Speaking before Hackett at the same conference today, Optus managing director of its wholesale and satellite divisions, Vicki Brady, confirmed the company was planning to offer wholesale NBN services.

“This is why Vicki Brady wants to sell wholesale services to 10,000-customer ISPs – because they’re dead without it; Quite deeply,” Hackett said.

In addition, the Internode chief said it wasn’t just smaller players who would be affected by the pricing structure that NBN Co has outlined.

Hackett pointed out that even ISPs like Internode – which has about 250,000 customers — would need to gradually migrate its customers onto the NBN fibre gradually over the next decade. This means that it would need to suffer the claimed high connection costs until it bulked enough economies of scale through customer connections.

“How many retail national service providers do you think the NBN is planning for?” Hackett asked. “The feeling of the rhetoric is that there wil be hundreds of them. It’s actually just five. We’re the number five player in the industry. Anyone smaller than us, never gets across the valley of death.”

NBN Co will be invited to respond to Hackett’s claims.

Video credit: Delimiter


  1. The irony of the accc killing future competition to satisfy existing fibre backhaul owners is pretty amusing, but surely it is still easier for ISPs than now?

    Correct me if I’m wrong but if an ISP wants to avoid Telstra wholesale they have to install their own dslam at not inconsiderable investment at each exchange.

    Am I missing something?

    Anyway as I’ve said before the rational for having lots of ISPs goes away once you have uniform pricing and service. The only reason they existed before was that ISPs would undercut Telstra’s 50%+ margins by cherry picking exchanges in particular areas, or bundling.

    • I think what Simon would say in response to this would be something along the lines that the current broadband environment has evolved and been regulated over a decade now — so ISPs can actually make money by installing their own DSLAMs and so on — it’s more or less stable, aside from some fluctuations from Telstra Wholesale and so on.

      In the new environment, this may not be the case — and NBN Co’s calculations about ISPs actually being able to afford its prices may be off.

      • I’m trying to decide if Simon has this right or wrong, and in the end I think it gets down to exactly what you’re trying to do as an ISP.

        Breaking it down, a single user connection is priced by the AVC price. The CVC price is a “global” price spread over all of your users – the larger your user base, the larger amount of CVC backhaul you’ll need. Exactly how much you end up buying for your users depends on what contention ratios you want to offer.

        The ISP that spends a little more to have “better” contention ratios than one that spends a little less to save dollars is likely to attract more customers through gaining a reputation for having “a faster network”. In the end, how much an ISP buys gets down to their own individual business plans, and has little to do with what NBN does in the overall scheme of things.

        After that it gets down to how many POIs you choose to service, and HOW you choose to service those POIs. Some ISPs will choose to have their own gear at each POI they want to service – (rackspace will be available for rent) – but this doesn’t have to be every POI.

        Your “equipment” – (as explained in the fibre session at the last industry forum) – might be as simple as a single crossover cable at those POIs you want to service.

        If you want to service the entire country, you’ll need at least that single crossover cable in each of the 122 POI locations. If you want to do basic internet data only, you’ll need 122 crossover cables – that’s all – NBN would probably do that for you anyway.

        If you’re a simple, basic ISP – and there are PLENTY with less than 2000 users – that’s all you’ll need on the user side.

        On the “ISP side”, back at your “home” POI, you’ll need to pay for a NNI (Network/Network Interface) to cross between the NBN and your internal data centre gear.

        Then it’s your upstream bandwidth, which is not handled by NBN in any way. These costs exist now, and will continue to do so. This market will remain competitive.

        Some upstream providers may use the NBN for backhaul, but as far as you’re concerned, it’s just a pipe out through a VLAN on your gear – as long as you’re getting your data and getting it reliably, you generally don’t care HOW they’re doing it.

        At the basic level – (just pure internet data) – I can’t see how the model changes in any way – for ISPs wanting to just data, it comes out about the same, and you can extend your geographical reach for peanuts.

        For a “grand scale” ISP, with IPTV, and VoIP, and gaming servers, and business services, etc, etc, the costs are going to be higher than a basic ISP, but once again, that’s the same as now.

        The more services you offer your customers, the more you can charge them. That’s the market in action, and what the government is trying to encourage.

        Offer great services, attract customers. Larger ISPs will always have an advantage – (particularly when doing deals for upstream bandwidth) – just as they have now.

        Smaller ISPs will have a chance to grow that they didn’t have before – I’m just puzzled where Simon is coming from on this one.

        • I will clarify one point above – probably shouldn’t have used the phrase “global price spread over all of your users” – I should have said “global price spread over all of your users ON A PARTICULAR POI”.

          Once again, thinking ahead and typing slower than my brain ;)

    • Would smaller service providers be locked into the traditional wholesale methods if they were looking to extend to nation wide coverage?

      In other industries there are examples of several smaller providers banding together to compete with larger players.

  2. … it would cost as little as $27.28 per month plus GST …

    That’s entry level (12M down, 1M up) and at about 50:1 contention ratio (so during peak periods you would be lucky to get 1Mbps down of actual data). It’s also “class 4″ bandwidth, which is the lowest priority but there’s some hint in the pricing that you might get enough class 1 to run a phonecall, it”s all a bit vague exactly what you get.

    …. forced to provide what NBN Co calls a Connectivity Virtual Circuit (CVC) connection to all of these locations in order to provide a national service…

    Is there a rule that all ISPs must provide a national service? If there is then that would be a new rule, perhaps part of recent legislation… seems like a strange rule to have.

  3. If you want to provide 100Mbps services then 200Mbps of CVC per POI is a good start, at 120 POIs that in itself is $600,000 per month.

    Now how about getting to those POIs, let’s say on average it costs $1k per POI to backhaul that data back to your Capital City POP to terminate, that’s another $120,000 per month.

    Some POI’s might be reached via a cross connect, others via dark fibre, others via managed ethernet, either way it’s a cost that will vary from small to large. I just picked $1k as an average.

    So we have ballpark $720K per month which an ISP wanting to provide national services will need to fork out each month to reach those POIs and have a minimal amount of CVC.

    For 100,000 services that’s averaged out to $7.20 on top of the AVC cost. Of course the users won’t be averaged so some locations you will need to purchase more CVC, adding to that cost.

    Then you have the monthly NNI Fees, your colo fees for a CBD POP, equipment costs and all this before you put a single bit of IP onto the service or provide some sort of support.

    Things to think about…

    • And I should point out….

      ISPs/RSPs will have to pay these new costs to connect to the NBN in addition to maintaining their existing infrastructure for servicing their existing customer base, while a migration process to the NBN over no fixed time takes place.

    • Of course the users won’t be averaged so some locations you will need to purchase more CVC, adding to that cost.

      So some will need more, but none will need less?

      • Correct.

        If you plan to offer 100Mbps services at a POI, as a minimum 150 to 200Mbps would be a minimum on the CVC to account for at least one service bursting to maximum line rate.

        At the locations with higher subscriber count you will purchase more CVC to match the desired contention or traffic profile, however there is definitely a minimum entry level cost of providing services.

        • If you plan to offer 100Mbps services at a POI

          But you have to offer 100Mbps at every POI?

          • Of course you don’t have to sell services greater than 12Mbps at a POI.

            I can’t imagine why anyone sane would restrict their product offering in such a way, most ISPs will want to have the top wholesale speed on offer at least as an option for their user base.

            Surely in each of the 120 POIs you’ll have a business or two that would want something more than 12/1Mbps.

          • But that’s the whole point, isn’t it? Just by changing assumptions, you arrive at two sets of numbers that are orders of magnitude different. Michael’s numbers assume you don’t offer more than 12Mbps at any POI. Yours assume you offer 100Mbps at every POI. The reality will probably be somewhere in between.

            And what about regional ISPs that have already carved out a customer base in a particular region. If all your customers are in Tamworth, you’re not going to be buying any CVC in Perth for instance.

          • For sure, I was assuming a national ISP, if you are going to cover a region to serve 1, 2 or 3 POIs then yes the numbers will clearly be different.

    • I’ve been tossing some numbers around inside a spreadsheet.

      If you look at 122 POIs (the current proposed number, it’s popped up by a couple in the last few weeks thanks to the ACCC massaging the numbers.

      250,000 customers spread across 122 POIs is 2049 customers on average, per POI.
      10,000 customers spread across 122 POIs is 82 customers on average, per POI.

      Assuming 12Mbps services – (we can only really compare at these speeds, since that’s the only speed available everywhere at the same price at this stage)…that’s:

      For the LARGE ISP:

      2049 customers at 12Mbps = 26639Mbps per POI (if all users running flat out).
      Reasonable contention ratio in the consumer market is about 50:1
      26639Mbps with contention ratio applied = 532Mbps per POI to supply via CVC.
      At $20 per megabit from NBN for CVC, that’s $10656 of CVC haulage per POI.

      $10656 x 122 POIs = $1300032 per month for CVC haulage costs.
      $200 pcm X 122 POIs = $24000 per month for NNI charges.
      250,000 x $24 AVC charge = $6,000,000 per month per AVC charges.

      TOTAL of $7,324,032 per month of charges from NBN Co.

      Cost per user (250,000) = $29.29

      For the SMALL ISP:

      82 customers at 12Mbps = 1066Mbps per POI (if all users running flat out).
      Reasonable contention ratio in the consumer market is about 50:1
      1066Mbps with contention ratio applied = 21Mbps per POI to supply via CVC.
      At $20 per megabit from NBN for CVC, that’s $426 of CVC haulage per POI.

      $426 x 122 POIs = $51972 per month for CVC haulage costs.
      $200 pcm X 122 POIs = $24000 per month for NNI charges.
      10,000 x $24 AVC charge = $240000 per month per AVC charges.

      TOTAL of $315,972 per month of charges from NBN Co.

      Cost per user (10,000) = $31.59

      That’s a difference of $2.30 per user, per month.

      • It appears your numbers don’t include getting the data from the POI to the ISPs POP, that’s another $100K plus as a minimum.

        In the small ISP numbers you are assuming that the ISP you aren’t going to sell a single service greater than 21Mbps at each POI, unlikely.

          • Who said it was the responsibility of NBN Co?

            It’s simply another cost input for ISPs (like IP traffic) that will make up the total cost of the service.

          • But it’s not a new cost, is it? I mean, they already have to pay to connect to Telstra’s exchanges today – and there’s a lot more exchanges than there will be POIs under the NBN.

          • It depends on the situation, for some it may be a new cost.

            e.g. for those selling Telstra Wholesale ADSL ports they would currently collect that data from an aggregated point in a CBD location or similar.

            So if they now want to take customers at the POI then it would be a new cost.

            It also depends on where the POIs are, it depends if existing backhaul in place to the site can be used, it depends if an ISP already has a presence at the exchange/POI.

            Point being NBN’s AVC+CVC numbers are not the be all and end all of the pricing game.

          • Sure – but the AVC/CVC component of the stack is the only part that is changing, so it is the only thing that’s relevant to the discussion, since it is only the NBN portion he is speaking about.

          • Exactly.

            Hackett is trying to make the point that moving from the existing DSLAM model to the NBN model makes it all more expensive, so that’s the portion of the network I did numbers for.

            He speaks of $106 for a connection – and that’s undoubtedly for a full 100/40Mbps connection with a shed load of data allowance – so he’ll probably have to charge $150 for it – so what? There are DSL packages around for that price RIGHT NOW.

            Storm in a tea cup – if this is really how it works out, why is he the ONLY one coming up with numbers like this?

          • So what conclusions can we draw?

            I think we can say that it would be difficult to offer different speed plans in different areas, but is it absolutely necessary for every ISP to offer 100 Mbps in an initial offering? Quite possibly some could even get away with 50Mbps or <=25 Mbps if it was attractive enough and they wanted to go it alone.

            More likely however, smaller ISPs are going to stick with the POI's where they already have a strong loyal subscriber base, and go through Optus/Telstra wholesale for the rest, or split their customers between wholesale and their own ports so they can offer 100 Mbps.

            Which brings me back to my initial thoughts, that if you were starting an ADSL based ISP from scratch and avoiding wholesale completely, you'd be faced with a massive investment that could only be undertaken if you were one of those 5 ISPs, and I think it's critical to note that a lot of ISPs and even Optus chose not to go down this path for some time. So yes it's only a problem if you thought as some initially did that the NBN was a complete reboot of the ISP industry.

          • If you were starting an ISP from scratch you would not go national on day 1. Of course, given that a new ISP is facing an existing saturated market, it would only be worth considering if you had some niche and could concentrate on excellent service within that niche.

            The minimum sized ISP would have only 12M bandwidth at CVC in only one POI and to achieve 50:1 contention on this they would need only 50 customers (in the same local area) all buying entry level plans. They would pay 12 * 20 = $240 per month in CVC fees and 50 * 24 = $1200 per month in PVC fees. They probably would find someone at that POI they can buy an upstream feed from, which at current rates, for a small party to buy 12M of upstream Internet would cost many thousands, I’d estimate around $4000 per month. Thus the NBN charge would not be their biggest problem.

            Let’s say total costs of 240 + 1200 + 4000 = $5440 per month, over 50 customers (with no profit, and no running expenses) comes to $109 per month.

            The problem for small ISPs being the upstream Internet feed, which is really the problem today… but no doubt that price will come down.

  4. I don’t think the NBN pricing is insane, what I think is insane is thoughts that any ‘meaningful’ ISP must have nationwide coverage along with thoughts that ISP’s with <10,000 customers do anything but purchase from existing wholesale providers currently. Providers with <10,000 customers simply don't have the numbers to build DSLAM's.

    iiNet started in a garage, look at them now – they sure as hell didn't build hundreds of DSLAM's from the word go.

  5. It seems Mr Hackett is still under the impression it’s one CVC per POI when in fact it is one CVC per CSA. It doesn’t matter how many POIs there are if the number of CSAs is the same. Back in the 14 POI days there were still ~200 CSAs, so ~200 CVCs were required to provide a national service. Guess how many CVCs are required if there are 120 POIs with ~200 CSAs. Yep … ~200. Just the same.

    • The CSA omission was our mistake, not his. The presentation is correct.

      Mr Hackett wants someone to give him a pony free. I can’t blame him for that, but it just ain’t going to happen, because someone does have to pay for this thing, and the CVC is the best way to do it.

      • Oh, so now we get an admission that the NBN will raise prices, only took half a year, but we got there!

        Now we just need Conroy and Quigley to say the same thing publicly to give an accurate picture, instead of somehow assuming that NBNCo can pull green notes out of their arse.

        Don’t see this happening any time soon though

        • The old Simon says rule applies when it suits, eh?

          Not so long ago (Feb 2011) Simon said he was pleasantly surprised at the 100Mbps NBN uptake in TAS, as it was higher than expected…

          But the Simon says rule didn’t seem to apply back then, as other naysayers said, oh of course he would say that. Sigh!

          • Tasmania is still under NBN trial, which has no wholesale cost whatsoever (AVC and CVC). There isn’t currently a NBN connection which uses the proposed pricing in the business case

          • Oh the “business case”.. dat ego?

            The same business case, that when suits the thread at any other given time, you you refer to a “toilet paper” (DON’T YOU?)… So that is the business case you now hold as your star witness?

            LOL…and I refer to your clone as Mr.Contradiction…OMG!

        • I must have missed the part of the presentation and discussion which said what the retail pricing is going to be, or at least what the ‘deteego unlimited’ plan is going to cost.

          • Deteego Unlimited Plan – only $38 Billion per month including delays and overspend at no extra cost.

            I dont think anyone would sign up for that one.

          • Are you asking me to give figures on behalf of other companies?

            The fact that the ISP’s (and others) are complaining, means that this rise in pricing is significant (one way or another)

      • I like this point in the PDF

        Slash CVC pricing (or just remove it)

        Pity NBNCo will not do that, since they won’t raise the capital they need in order to pay off the network, even with a just an increase in AVC from $24 to $29. Run through the numbers and NBN will not get the revenue with Simon’s proposed changes

        As Simon Hackett correctly pointed out, the pricing structure of NBN is bringing us back to the old dark ages of Telstra’s Wholesale monopoly. Those days were fun, horrible quotas and only a few big ISP’s

        We’ve drunk our poison, have you had your fill?

        • This discussion has only been pricing entry level plans. Work out the quota on an entry level plan:

          ( 12Mbps / 8 ) * ( 30 * 24 * 60 * 60 ) * ( reasonable utilisation estimate ) / ( contention ratio )

          For 50:1 contention and a utilisation of 50% you achieve a monthly quota of about 40G bytes.

          That’s hardly an exciting plan, worse than many mid-level ADSL plans right now, so the entry level plan needs to be retail priced at the low end to even be competitive with ADSL.

          I saw Jeremy running round telling people they were going to get 100Mbps continuous (i.e. uncontended) which of course won’t ever happen but if it did happen, the CVC price on 100Mbps would be $2000 per month. Which is a different story.

          For NBN to make financial targets, their AVERAGE user needs to be somewhere around $45 per month so every $27 per month plan they sell must be balanced by at least one $63 per month plan (or they have to sell a few massive business plans).

          • Oh of course, I agree with you, this is blatantly obvious, and it doesn’t have people going around doing deliberate misinterpretations of CVC charges by using the extremely light plans as examples and putting them forward as average or normal case scenarios.

            I have said this months ago, it seems the ISP’s are talking out at a time when they get the most attention (which is good).

            The CVC charges get quite spastic when you account for internet plans that aren’t trivial in usage or download behavior

  6. Riiiight, so how many “national” ISPs have fewer than 10,000 customers?
    Internode is running down the value of smaller players so it can buy their customers.

    • I don’t know if it’s his real motive, but one of my first thoughts when I read this whole article was it will scare people away from signing with smaller ISPs towards the larger ones.

      On a connection by connection basis, there is very little between what a “small” ISP would be charged by NBN and what a “large” ISP would be charged.

      Internode – like any business – looks at its own business plan and makes decisions/statements/assertions based on where they are, and where they’d like to be. Any ISP with 10,000 users – (before or after the NBN) is not going to be shooting for an ISP with 250,000 users.

      They’d lose everytime.

      Same now, same after.

      • @MichealWyres

        I agree, the wholesale ADSL market at the moment is more diversified and under closer cost control by those ISP’s/Telco’s that have their own DSLAM’s, and are not reliant on Telstra Wholesale for everything they sell.

        Small ISP’s today can mix and match from a number of wholesalers in providing their retail plan offering, with a mix of ADSL/ADSL2+, line rental, PSTN , ULL, LSS etc, post NBN they are 100% reliant on the NBN Co and a CVC wholesaler for every fixed line product they sell.

        Internode has expressed concern about 10,000 customers and less ISP’s, I would also be concerned if I was a major DSLAM roller also, no more Naked DSL a major competitive edge over No 1 BigPond and no more tweaking technically and on price of your plans because you own/lease exchange gear.

        Without their own DSLAM’s and cost control how will ISP’s differentiate from what BigPond and Optus offer? – a price war against those two they could not sustain?

  7. The first five ISP’s includes Internode at five? err what happened to TPG with about 500,000 customers, that makes Internode sixth.

  8. All of this controversy about the selling of CVC circuits pricing being too high is based on speculative price modelling on what a NBN wholesaler will charge.

    So far Optus has announced it will be a NBN wholesaler, no doubt Telstra will also at some point and maybe others with well established infrastructure like iiNet.

    If those wholesalers are competing for business from the 10k customers or less group of ISP’s I am sure they will compete with each other on price and the value add products they add to the package to sweeten the deal.

    • The issue is that this pricing adds another margin and moulds the market to benefit certain types of products and ISP’s. The value added packages would have to be a massive rip off to cover the extra cost borne by ISP’s, either that or basic internet prices will basically increase

      Competition only works if that segment of the market is “free”, this isn’t the case.

      • @deeteego

        I agree with you that with more players between the NBN Co ‘raw’ wholesale product and what the punter pays for a retail plan there are the higher the end plan price will be.

        Does this mean that Telstra and Optus for example will always be able to undercut smaller/medium ISP’s because they don’t have to buy through the ‘margin added’ CVC route?

        This reminds me of a complaint ISP’s had against Telstra in the past with the ACCC case where BigPond was selling a plan cheaper than what ISP’s could get it off Telstra Wholesale for, and Telstra was fined for that, and it led to major drops in TW ADSL1 wholesale pricing.

        It seems with the NBN (in theory anyway) the more things change the more things stay the same, especially when it comes to market dominance of the big two at the BB fixed line level.

        • At the end of the day, with Telstra ISP’s use ULL/LSS wherever they can (which is the vast majority of areas) has a flat $16 apart from very remote areas. All the other costs are on the ISP’s side, which they can either cover themselves or with other companies (this is “free” market competition)

          As an example, for both local and national connections, ISP’s can use ISP, which if you check their site (http://www.pipenetworks.com/), PIPE charges a flat fee, regardless of how much contention or data you are sending through their fiber. Other ISP’s make their own local/national fiber network (such as Agile), in which case the cost for ISP’s is just an initial capital cost (after when its payed off, its basically “free”). Companies such as Exetel completely run off other companies from the other spectrum. The conclusion is that because everything apart from the physical copper connection is left to the market, the costs in those areas are basically due to competition, there are no set prices or monopolies. When Michael says there are costs to local backhaul, he is correct, but he is vastly overstating/exaggerating what the costs are (this is his deliberate misrepresentation of the matter, and the fact that his $2.50 CVC per month figure is out by factors demonstrates this)

          Under the NBN, this whole things flops on itself. The prices that NBNCo charges are set in stone, no ISP can lower the costs and since its a monopoly and off another company. In basic business sense, this is basically just another margin on their costs, one which they cannot reduce, so it will push up prices. Of course this was obvious ever since NBNCo released the business case, these CVC costs are needed for NBNCo to pay off its network, and NBNCo’s costs are factors higher then Telstra’s fixed line network.

          So, finally, internet price costs will rise, and not by a trivial amount for the non light users. The CVC charges are framed in such a way that the heavy downloaders will bear most of the costs, which the ISP’s claim here (http://www.smartcompany.com.au/internet/20110323-nbn-wholesale-pricing-model-threatens-unlimited-plans-telcos-warn.html)

  9. “The more things change, the more the stay the same”…alain?

    Except, with the NBN…

    The biggest player (Telstra) don’t own the network/monopoly last mile, leaving their competitors at their mercy!

    The NBN will provide for those who do not currently have (those you naysayers pretend don’t exist).

    The copper (which will not last forever) and certainly will NOT be able to handle the traditional technological advancements in years to come, will be replaced and we WILL be future proofed, as much as is currently possible. As such, the words of Bob Katter, referring to the opposition (and you naysayers too?) are reverberating clearly …‘They think we should wait for some science fiction fantasy to jump out from behind a bush, (but) we’ve got an offer on the table, and we’re going to take it.’ LOL…priceless and spot on, Bob (never thought I’d say that…)!

    And many more positives beyond the status quo, we have already gone over and over, which you simply ignore, pretend will not be advantageous or are sadly unable to comprehend will be advantageous.

  10. Yes well, I was feeling a bit crook earlier. Jasmcd..LOL

    And for the record, the comment I agreed to was the previous one ( a few comments above where it currently sits) which is a bit deceptive…!

    But I have absolutely no reason to disagree with these guys when they post comments which aren’t typical baseless FUD.

  11. The NBN is becoming a clown act every day.
    I like competition it adds choices to the consumer if the NBN is going to be another Telstra then the consumers are going to get screwed.

    We need the NBN we don’t need another monopoly and we don’t need the censorship plans either.

  12. “This is why Vicki Brady wants to sell wholesale services to 10,000-customer ISPs – because they’re dead without it; Quite deeply,” Hackett said.”

    Well yeah they probably want to sell it to ISP’s like you as well, just like they sell wholesale Naked and ADSL2+ services to you today.

    The point is, are you interested if the price is right?


  13. Hi Renai,

    I’m having a problem with this entire debate. Simply stated, I don’t understand the situation well enough to establish a clear picture of what people are arguing about.

    I’ve been in the IT industry for 34 years now and have worked in or managed virtually every discipline that’s been invented in that time. While I no longer write software, I do know how computers work and how information systems are designed and built. I do have some idea of how data communications work too. When I need to, I can generally find something that gives me a clear explanation of any technology I need to understand, so that I can make informed judgements.

    But in the case of all this argy-bargy about NBN interconnects and so on, I am all at sea. Yes, maybe I could find something that will give me a clearer picture, but then I wonder how many others watching and wondering about this debate are in the same situation – watching people having heated arguments about stuff we can’t even visualise.

    Now think of the business leaders out there. Not only do they not have any idea what this great white elephant (in their eyes) is all about, but they can quickly and reasonable form the view that the technicians also don’t know what it’s all about because a) they are all arguing and b) what they are arguing about is unfathomable.

    This is one of the classical problems with governing IT – so much of the discussion and debate is conducted at the deep technical layers, when it should be lifted to a higher level so that many more stakeholders can grasp the issues, form an opinion, and give voice to their views. So that we can, as a nation and a community, properly govern the development of the NBN, we must make the debates more accessible to the ultimate stakeholders and not just to the technical players in the inner circle.

    So can you please consider preparing, or getting somebody appropriate to prepare a “Lay persons guide to the NBN” that helps us understand what the system is likely to look like, and what the various items of contention are all about? I’m sure that many business leaders, and more than a few technically literate people would love to have access to such a resource – and if we had it maybe more of us could make an informed contribution to the debate…

    The guide needs to overcome some challenges – it needs to be simple and clear, in plain English with a minimum of technobabble and clear explanation of the unavoidable bits of babble. It needs to help us all understand what people are talking about when they bang on about points of interconnect and virtual circuits. Most importantly, it needs to do all this without propaganda – we don’t need more on how wonderful it will be – we just need something that tells us, in a way we can understand, how it all works. It needs to help us bring the debate out of the mysterious realm of technology, to a level where competent thinkers can make informed and balanced comment. Oh and it does need to show us how ISPs and other telcos fit into the picture – overlaying the business model on the technology model.

    It’s a tall order, but it can be done. The fact that a Google search for NBN Overview does not return such a document in the first four pages of results gives a fairly good confirmation that no such guide exists right now.

    • Contention is basically a term that describes how much “stress” a part of the network is on. If you for example have lots of people heavily downloading at one time, then the contention is very high, where as if you have light users that burst at different times of the day, then the contention is low. Its the same for other utilities, if you have a lot of people using electricity constantly at the same time, that causes high contention. Of course the difference between other utilities and internet, is that the actual cost of sending data through a pipe is basically nothing. This is not the same as water for example, since water (which is what you are sending down a pipe) is an expensive resource

      The NBN uses a dual pricing model (which is obviously going to be forced, since its a fixed line monopoly). One part of the cost is the AVC (Aggregated Virtual Circuit) which has a base cost of $24 (for a service of 12mbit) and increasing in price with tiers (for up to 1000/400). From the RSP side, this is a cost per connection to premise

      The CVC is another cost, which works like a resource. RSP’s purchase allocations of CVC (in $20 blocks iirc) and as users download from the internet, depending on how much contention they create the ISP’s then use that allocation to pay it off. In terms of globally around the world, very few companies or corporations have CVC charges, let alone mandated monopolies. ISP’s, pre 2006 were forced to use this pricing model (AGVC) when you had to go through Telstra Wholesale for internet, back then Australia was notoriously noted for having incredibly low quotas and high internet prices. The reason why is obvious, the CVC is a charge that is basically proportionate to how much your users download. The more your users download, and more importantly how fast they download (especially at once during peak hours), the more ISP’s get charged with CVC. The CVC charge is structured in such a way that heavy downloaders are heavily penalized (where as light users won’t see much of a difference). Of course heavy downloading is the whole point of the NBN

      In regards to what is currently paid by ISP’s/RSP’s, most ISP’s/RSP’s use ULL/LLS service (its available almost everywhere apart from regional areas). The ULL is a flat $16 a month, LSS is ULL + $2 + line rental (for phone). The ULL/LLS is the only costs that ISP’s are forced to pay to Telstra, all other costs are governed by the free market. There are no CVC charges, just a flat connection to the physical copper line (which is the ULL). In terms of other costs for ISP’s (which you won’t have to do with NBN) there is infrastructure such as DSLAMS, and local backhaul (laying fiber links). Both of these costs however are initial capital costs, and once paid off they cost the ISP’s “nothing” (ISP’s also have a choice to use other companies for this infrastructure, i.e. market). The only charge for actual data usage that is forced today is basically downloading from international links (something that ISP’s will have to pay on NBN as well, this won’t change). Companies such as TPG, that own international links (such as PIPE) have this cost basically free as well (one of the reasons why TPG offers unlimited for $60, they also own the national and local fiber links)

      So basically under NBN, ISP’s have to pay for two charges based on bandwidth, a contention charge (CVC) and international links. Currently, ISP’s only pay for international links.

      If this sounds expensive to you, then yes you are right, and the basic business reason for this is that NBNCo has to pay off the entire capital for building the network, which is ~36 billion dollars +interest (this is just the CAPEX, NBN will also have OPEX). The current CAN network already had its capital paid off long ago, it only has an OPEX to pay for. If NBNCo didn’t have to pay for the capital (and wasn’t expected to be financially viable) then CVC charges wouldn’t even need to exist. The capital costs are so ridiculous basically for the reason that to install FTTH, you have to install cabling onto everyones back yard, which is incredibly expensive due to the labor

    • http://en.wikipedia.org/wiki/Network_congestion




      The trouble with the stakeholders meetings is that those stakeholders have to show enough responsibility to bring themselves up to speed with the issues. If I “educate” them with my point of view then they are not bringing anything new to the meeting — just carbon copies of my ideas. All the basics are easily available but putting it together requires effort, no way to bypass that.

      As for the complexity of NBN pricing, that’s partly a result of them trying to be all things to all people. They need a complex system of prices to cover the wide variety of use cases. It’s also a result of them being a political entity wanting to release information to the public in a way that allows them to say, “Hey Telstra ULL is only $16 per month, but our plans are only $24 per month so hardly any change.”

      Thing is, we have $36 billion investment (or some rubbery figure) at 7% interest which comes to approx $200 million per month (or $7 million a day if you prefer) that needs to come from somewhere. All the rest of the discussion is just shuffling around working out who gets to pay (… given that NBN is going ahead, one way or another, somehow the Australian people WILL be paying that).

  14. Why does all seem obvious to me and something that expected, and why I disagreed with the ACCC’s high number of POIs in the first place.

    • I think the numbers will surprise you. I’ll repeat my above costing example using the large ISP numbers, and a reduction from122 to 14 POIs:

      250,000 equals 17857 users per POI – (when POIs reduced to 14)

      17857 customers at 12Mbps = 214284Mbps per POI (if all users running flat out).
      Reasonable contention ratio in the consumer market is about 50:1
      26639Mbps with contention ratio applied = 4285Mbps per POI to supply via CVC.
      At $20 per megabit from NBN for CVC, that’s $85700 of CVC haulage per POI.

      $85700 x 14 POIs = $1199800 per month for CVC haulage costs.
      $200 pcm X 14 POIs = $2800 per month for NNI charges.
      250,000 x $24 AVC charge = $6,000,000 per month per AVC charges.

      TOTAL of $7,202,600 per month of charges from NBN Co.

      Cost per user (250,000) = $28.81

      122 POI version = $29.29
      14 POI version = $28.81

      In the case of these numbers, the difference between the 122 POI model, and the 14 POI model is a grand total of $0.48.

      The beauty of the 122 POI model is that it allows carriers to be more specific about which areas they wish to provide coverage to, which is why the ACCC sought to increase the number.

      • oops!

        The line “26639Mbps with contention ratio applied = 4285Mbps per POI to supply via CVC.“…

        …should read “214824Mbps with contention ratio applied = 4285Mbps per POI to supply via CVC.

        Lazy cut and paste on a Friday… ;)

      • Hi Michael,
        You really also need to take into account the implications of long hauling 122 interconnects back to a POP in a capital city and the hardware costs of 122 NBN interfaces versus 14.
        A single Nextgen protected path BNE-SYD GigE Link costs ~ 10-15K/month, and NN are one of the cheapest backhaul carrier.
        122 x $10K = $1.2 million per month, just for the backhaul. Granted, not all of the POIs will be as long as BNE-SYD, but some will be even longer, depending on where your POP(s) are located.
        Even if you halve the backhaul cost to $500K/month, for an ISP with 50,000 users that still adds $10.10 inc GST to the monthly tail cost.
        Also, I would dispute your 50:1 contention ratios. That is fine for vbr-nrt traffic, but one of your main arguments for the NBN is that people will be using it to stream video. If everyone turns on the TV to watch the 6-7 news on various channels, then that 50:1 is shot to bits. I’m sure you know that you need to build for peak, not average.

        • The CVC actually IS the haulage back to the POI you connect your POP(s) to.

          You’re assumption is that you have to have a link to every POI. That’s false. A POI is connection between the core network ring of the NBN and the distribution network of the NBN in each area. You only have to connect your network to a single POI. That connects you to the core ring of the network, POTENTIALLY allowing access to each and every POI in the system. Only if you choose to service customers connected to a particular POI will you need to be able to pass traffic through it.

          You would probably connect your POP(s) to a few of them, in case a POI goes down.

          Let’s say you have three. $15K each? Sounds reasonable. $45K a month. Spread over 250,000 users? That’s $0.18 a user.

          Even if there were 122 of them – (15K x 122 = $1,830.000) – across 250,000 users, that’s $7.32 per user. But the thing is, POI to POP haulage is NOT the responsibility of NBN Co, and this cost exists for ISPs now. There is absolutely no change here.

          They are paying your $15K for links now, and they will still need to do so.

          Hardware costs to get from the POI to the distribution network? As explained in the fibre session at the most recent industry briefing – can be as simple as a crossover cable. How much does a crossover cable cost? You can have more equipment in place depending on what services you want to offer your customers, but if you’re offering them more services, you get to charge them more for it!

          CVCs can also burst for no extra cost for unexpected trafffic peaks.

          Contention ratios – standard rates for consumer level, best-effort routing data is generally around that mark. Some ISPs go lower, even a lot lower. But the lower you go, the more bandwidth you are buying, and the more you have to charge the customers. Every ISP has to balance their contention ratio between costs to buy, charge to customers, and their own expected profit margins.

          Nextly, CVC pricing – there are several tiers of packet priority with the NBN.

          REALTIME (highest)
          BEST-EFFORT (lowest)

          Packets tagged “REALTIME” (set aside for voice) are routed first, followed by “INTERACTIVE” (IPTV, HD video conferencing), TRANSACTIONAL (set aside for business data), and BEST-EFFORT (everything else – web browsing, email, etc). BEST-EFFORT is charged at $20/Mbps

          CVC pricing is tiered against these priorities. If you want a channel at the “INTERACTIVE” priority, it will cost you more per megabit. That’s only sensible – you’re seeking a higher level of service from the network, so you get charged more for it, and once again, you can charge the customer more for it.

          As for my numbers, my example only compares like-for-like services (BEST-EFFORT, 12Mbps, basic data services) – to demonstrate that their is no significant difference between a 122 POI model and an 14 POI model, when comparing like-for-like services.

          If you do the numbers, you’ll see that the difference remains equally insignificant all the way up the range of speed increments. For example, if you compare like-for-like at the 100Mbps speed, for an ISP with 250,000 users, the cost per connection works out at $77.43, and for an ISP with 10,000 users works out at $79.71.

          It is completely pointless comparing an entry level, data-only 12/1Mbps service, with an all-singing-all-dancing 1000/400Gbps plan tagged INTERACTIVE for routing purposes.

          They are not the same, and are not priced the same.

          Now, if we want to relate back to my Foxtel example – (might have been in another thread) – an NBN Co NTU in/on your house can serve up six separate and completely discrete services.

          Let’s say you sign up for an Internode basic internet service at 50/20Mbps, with 50GB of data a month, you might get charged $65 for it. That’s gonna be set up at the BEST-EFFORT, $20/Mbps CVC tier within the network.

          You get Foxtel connected, and they provision A COMPLETELY DIFFERENT CIRCUIT into your house. Say you have two STBs, needing about 12Mbps each, so they might provision a 30Mbps service at the INTERACTIVE tier of CVC pricing. This might be $65 also – lower speed, but similar cost due to it being at the higher CVC tier.

          This is the stuff that 99.9% of people DO NOT UNDERSTAND about the NBN. The Coalition won’t even discuss the details because to them it’s all about the money. The government just don’t know how to sell it to the people.

          People should start turning up at NBN briefing sessions and listening to how NBN are building this network, instead of listening to dicks like Alan Jones and the Australian newspaper, who probably couldn’t turn a computer on, let alone understand how a network of this scale is built and operated.

          • “You’re assumption is that you have to have a link to every POI. That’s false. A POI is connection between the core network ring of the NBN and the distribution network of the NBN in each area. You only have to connect your network to a single POI.”

            OK, without semantics now, you are claiming 100% than an ISP can have 1 (one) single cross-connect to a single NBN POI and service NBN customers nationally from that one inter-connect??

            PS Real classy with the Allan Jones comments. I have never seen you post to the Australian Network Operators Group or at one of their annual meetings, so I could question your standing in the Australian network operators community, but I won’t stoop that low.

          • OK, without semantics now, you are claiming 100% than an ISP can have 1 (one) single cross-connect to a single NBN POI and service NBN customers nationally from that one inter-connect??


            They’d be stupid to do so – (one cut link between them and that single POP would bring them down nationally) – but it would “work”, absolutely.

            Highly contended – absolutely. But it would work.

            But if you’re an ISP with 250,000 customers, you’d have rocks in your head not to have more points of access to the core network. But you don’t HAVE to.

            If you’re an ISP with only 10,000 customers, you’re gonna get away with a lot less.

          • Note the use of the phrase “national RSP”.

            The big guys probably will connect to most, if not all – if for no other reason to be able to compete with each other in the capacity stakes. The competition at the big end of town is massive. Look at the big deal iiNet is making about being number two in the DSL space.

            As NBN Co suggest, Hackett is assuming that everyone wants to compete nationally – which is a massive assumption to make. If I’m a small ISP, I’m only going to purchase access to the POIs I need to access to service my customers.

            But he’s arguing OUTSIDE of the NBN. The connection from POPs to POIs is not part of the NBN. They are connections that they have to implement now, to all the exchanges. I bet he’s got more than 122 right now.

            As Simon himself said on Whirlpool late last year, when asked how Internode justifies where to put their own DSLAMs in:


            Positive business case (financially viable backhaul available plus a combination of enough existing customers and/or sufficient catchment to be confident of obtaining sufficient)

            (assuming he means “sufficient return” at the end there)

            So in October last year, it’s all about only putting his own DSLAM in an area where he’s got enough customers and confidence that he’ll make a decent return on it, yet now he’s saying that everyone HAS to be connected to all POIs?

            He seems to have either changed his mind in six months – (he’s allowed to) – or it’s one rule for him, and another for everyone else.

            I have a lot of respect for Simon and Internode, but I think he’s barking up the wrong tree here.

          • As for the “Australian Network Operators Group” – (despite having actually never heard of them) – I’m sure they exist, and do good work.

            But just because I don’t travel in the same circles as your good self, doesn’t mean you’re in any way in a superior position to comment, or that my knowledge is any less (or greater) than yours. I’m not driven in this debate by ego.

            I’m focussed on the points of the debate, not on myself.

  15. Thanks to deteego for the effort to explain the NBN. Unfortunately, even though I have delved into the internals of operating systems in the past, could once read BCD (Binary Coded Decimal) paper tape by eye at about two bytes per second, and have worked with electronics engineers to design and build special-purpose computers and peripherals, I am still utterly lost by the barrage of unexplained acronyms that are used in all sorts of discussions around the NBN.

    Michael Wyres is getting me somewhat closer – but there’s a long way to go yet.

    If I’m having trouble with this, imagine how difficult it is for business leaders and operators who just want technology that works properly at a fair cost – the message they are getting at present is that the NBN will be yet another technology nightmare in which those few who understand it plunder those who can’t get to grips with the minute detail.

    The Australian IT marketplace – especially the customers – desperately need a wise and skilled translator who can help us all understand what these “insert interesting descriptive word here” people are arguing about!

    Does anybody know such a person?

  16. Does anyone know how Simon Hackett intends to slash the CVC charges ?
    Is he suggesting the government forgo the return ?

    • He’s suggesting that they either:

      – raise the AVC price and reduce the CVC price
      – raise the AVC price and drop the CVC pricing all together

      …and he is quite right that doing either could/would end up delivering the same financial outcome for NBN Co.

      In my opinion however, more POIs makes the network more flexible in regards to how you service your customers.

      For the record, and in deference to Simon’s argument, I do believe that 122 is too many POIs, but I also believe 14 is far too few. I would have thought 66 was the correct number, to match the current number of national CCAs – (Call Collection Areas) – which would closely match much of the existing significant infrastructure.

      Rightly or wrongly, the ACCC decided to push it further.

      I see a lot or merit in Simon’s argument, but I see the same in the ACCCs – and I believe their position is “more right”. Only in the fullness of time will we see how it pans out.

      • So in both cases an increase in the AVC
        I guess thats going to make the entry level plans more expensive though. If you only want for example a 12Mbps 20GB plan, the increase in AVC will put that price up. I think NBN wants to keep those entry level plan prices similar to ADSL.

        The ACCC decision seems to be the main cause of concern
        Your suggestion of 66 probably would have been better

        I wonder if Internode themselves would become an aggregator ?

        Thanks for your comments
        You seem to understand this better than most

        • Do they become an aggregator? Excellent question – I guess that depends on their business plan, and which direction that they want to go.

          It’s a potential revenue stream for any ISP who thinks that they can make a go of that market, and there will be a market for it.

          NBN Co have always stated that they believe their model of being a Layer 2 wholesaler will encourage the expansion of the Layer 3 wholesale market, using their Layer 2 bitstream products.

  17. Ok, so a draconian government owned monopoly NBN Co legislates to eliminate (and compensate to the tune of $billions) existing infrastructure (copper, HFC, whatever) to avoid competiting technologies then comes up with a national pricing system that means only the top 5 ISPs will survive. Right.

    How many torpedoes does the whole NBN FTTP concept fire into the side of free enterprise and competition?

    I suppose they’ll be wanting us to queue up for bread next.

    • Your spiel would sound more believable and less politically biased/motivated without the colourful rhetoric, such as draconian and queueing up for bread…

Comments are closed.