$20,000 for 1Gbps: Turnbull’s Lateline lie

139

turnbull2

news Malcolm Turnbull appears to have made a deliberate attempt to mislead the public about the cost of connecting to the National Broadband Network’s upcoming 1Gbps fibre service, claiming on national television last night that such connections would cost “at least $20,000” a month, despite the fact that the Shadow Communications Minister is aware the cost is likely to be much less.

Last night, the Liberal MP squared off against Communications Minister Anthony Albanese in an election debate on the ABC’s late night current affairs show Lateline. Among other issues, Turnbull was asked by Lateline host Emma Alberici whether the Coalition’s planned fibre on demand service would allow speeds of 1Gbps. Currently, Labor’s fibre-based NBN infrastructure only allows speeds of up to 100Mbps, but NBN Co has stated that it plans to launch 1Gbps speeds later this year.

“[Albanese] said that fibre to the premises can deliver one gigabit per second, 1,000 megs, and you’re quite right, it can,” Turnbull replied. “Do you know what it would cost to have a guaranteed one gig’ service? At least $20,000 a month. $20,000 a month in combined virtual circuit charges … The reality is this: if you want to have a guaranteed one gig service, your retail service provider will have to buy one gig of CVC for you and that is gonna cost $20,000 a month.”

“For the average household?” asked Alberici. Turnbull responded: “Well, for any household, which is why, by the way, nobody will buy it other than businesses that need a very big …”

The Shadow Communications Minister explained during the conversation that retail Internet service providers such as Telstra or Optus, who would be the ones actually selling NBN Co’s 1Gbps broadband services to the public, needed to buy what is called a ‘connectivity virtual circuit’ to link their customers’ broadband access back to the wider Internet.

Turnbull’s argument that 1Gbps NBN services will cost $20,000 per month appears to be based on the argument that ISPs will need to provision a 1Gbps CVC circuit in order to provide guaranteed speeds to retail 1Gbps services.

In a post on Delimiter in June, iiNet chief technology officer John Lindsay noted that the cost of such CVC circuits from NBN Co would normally be $20,000 for a gigabit connection per month. Normally, that circuit would be shared between several customers, because it is not likely that each customer would need the 1Gbps speeds continuously ; for example, 10 customers, all with 1Gbps theoretically connections may share a 1Gbps CVC circuit, meaning that the cost of the CVC circuit to each individual customer would be only $2,000 per month.

This so-called ‘contention ratio’ formula — 10 to 1 in the above example — represents normal practice when it comes to providing any form of broadband in Australia. Contention ratios of between 10:1 and 20:1, for example, are common in the provision of ADSL broadband, while higher contention ratios may be used by cut-rate ISPs in order to save money and provide customers with a slightly degraded quality of service compared with ISPs which charge higher charges. Contention ratios of 1:1 are only used for guaranteed very high-end corporate broadband services and are not expected to be used for retail broadband services under the NBN, or even the vast majority of businesses. Only a very small number of broadband customers, typically large corporations, will require such contention ratios.

There is also another charge with relation to broadband services — what is called the access virtual circuit or AVC. This charge relates to the monthly access costs which ISPs pay per customer to connect customers to the NBN. Taken together, these two costs — the AVC and CVC charges — represent what it will cost retail ISPs like Telstra, Optus, TPG and iiNet to connect NBN customers.

In the case of the 1Gbps service, NBN Co has announced (PDF) that it will charge retail ISPs $150 per month in terms of its AVC costs. Taken with its CVC pricing, this means that retail customers, both businesses and consumers, could conservatively expect to pay in the realm of $1,200 to $2,500 for a 1Gbps broadband service, based on contention ratios of between 10:1 and 20:1. It is also possible, given the high quality of the fibre to the premises network which NBN Co is currently building, that higher contention ratios may be possible, which would drive down the monthly cost of 1Gbps further, below $1,000. These are estimates.

It is believed that Turnbull is aware of these charges and of the detailed cost modelling which NBN Co has made public with regard to its wholesale costs. The Liberal MP has extensively discussed NBN Co’s wholesale financial modelling in public on several occasions. In addition, the Liberal MP’s awareness of the $20,000 CVC cost for wholesale NBN access, and his use of the term “guaranteed” and discussion of contention in the Lateline debate indicate that Turnbull is aware that the $20,000 cost represents the extreme high end, and not likely real-world charges, of what retail NBN customers may charge for 1Gbps services.

Because of this, and because Turnbull directly stated that “any household” will need to pay the $20,000 monthly costs under the NBN, and the primarily residential context of the segment of the Lateline discussion in which the comments were made it appears that the Shadow Communications Minister attempted to deliberately misrepresent the cost of connecting to NBN Co’s 1Gbps service.

The news comes as both sides of politics continue to make attempts to mislead the public when it comes to each other’s NBN policies, ahead of the upcoming Federal Election. Last week, for example, Communications Minister Anthony Albanese appeared to have issued a media release deliberately misleading Newcastle residents about how the Coalition’s rival NBN policy would affect the area, with the Labor MP falsely stating that the NSW city would “miss out” on upgraded broadband entirely under the Coalition’s plan.

Labor MPs in general are also engaging in misrepresentation when it comes to the Coalition’s NBN policy. A number of ALP election advertisements have inaccurately claimed, for example, that Liberal policy would see Australians forced to pay up to $5,000, or else they would be left “on the old, slow copper network”, while connection to Labor’s fibre-based NBN would be free.

However, the Coalition has also made a number of misleading statements about Labor’s NBN project over the past several years. In one of the more blatant examples of misleading commentary, Federal Shadow Education Minister Christopher Pyne inaccurately claimed on national radio in October that the National Broadband Network had not connected any customers at speeds of 100Mbps, despite the fact that in fact, 44 percent of NBN customers connected to the project’s fibre infrastructure at that point had taken up such speeds. There have been several dozen other similar examples over that period.

In a more recent example, several weeks ago Opposition Leader Tony Abbott inaccurately claimed that the rollout of Labor’s National Broadband Network in Tasmania will take “80 years” to complete, in what Labor’s Regional Communications Minister Sharon Bird immediately labelled a deliberate attempt to deceive residents and businesses in the state.

Similar to the misleading infographics distributed by Labor MPs over the past several months, an infographic currently published on the Facebook page of the Liberal Party of Australia misrepresents Labor’s policy. It conflates Labor’s initial, $4.7 billion policy outlined in 2007 with its reformed 2009 policy, falsely alleging a blowout from $4.7 billion to $90 billion in the project, and a decade-long project timetable extension.

Malcolm Turnbull has been invited to respond to the statement that his claim on Lateline last night was a deliberate attempt to mislead the public about the residential cost of connecting to the NBN at 1Gbps.

opinion/analysis
There were a number of misleading statements made by both sides in last night’s Lateline debate. I started making a list as I was watching it, but then gave up as both sides just made too many little deceptions to keep track of. Most of what was said was true, of course, but I’d say something like between five and ten percent of the statements made could be substantially challenged in one way or another. I am aware that both Albanese and Turnbull were a little taken aback at times by the questionable statements each were making.

However, of all the statements made last night, Turnbull’s claim about 1Gbps services costing $20,000 per month is the most false. His use of the phrase “guaranteed” to fit in the deception over 1:1 contention ratios, which you almost never see in the real world, was a nice little trick, but the $20,000 figure was rightly questioned immediately by the host, Emma Alberici, and I laughed out loud when I heard Turnbull claim that “any household” would need to pay that cost. If that was truly the case, then “any household” would also need to be paying exorbitant fees ranging from several hundred of dollars into the thousands for 1:1 “guaranteed” contention ratios under Turnbull’s fibre to the node plan. Clearly that won’t be the case.

$20,000 per month for a 1Gbps residential NBN fibre connection? Nope. That just ain’t going to happen.

Image credit: Office of Malcolm Turnbull

139 COMMENTS

  1. The great thing about CVC costs is that they’re designed to bring in a certain amount of revenue.

    Unlike a commercial venture, they only need to make a profit 7% pa, and no more – so when users start using more data, and people start going for the faster AVCs, the price of CVC will reduce.

    If the NBN stays FTTP, I expect that in 10 years the population will be using a lot more data and many more people will be on 50 and 100Mbps+ NBN plans than the (appropriately very conservative) estimates in the NBN Co corporate plan, which will mean that the price for gigabit connections will drop to be easily affordable levels.

  2. $20,000 a month would be about right* in today’s market for a (contended) 1Gbps service, but the availability is rather… limited.

    One would expect that the future 1Gbps services on the NBN will cost about as much as a high-end 100Mb Ethernet service in the CBD; you will just get 10x the capacity down and 4x up.

    Oh, and it will be ubiquitous.

    *http://www.exetel.com.au/corporate-midband-ethernet.php

  3. I still have a problem with the claim that
    “that Liberal policy would see Australians forced to pay up to $5,000, or else they would be left “on the old, slow copper network”, while connection to Labor’s fibre-based NBN would be free.”
    is untrue.

    • That’s nice, well keep your opinion to yourself, because we’re sure as hell not debating that issue in this article. Any comments raising it again will be deleted without notice.

    • “that Liberal policy would see Australians forced to pay up to $5,000, or else they would be left “on the old, slow copper network”, while connection to Labor’s fibre-based NBN would be free.”

      If you do not pay extra under the coalition plan, you will have a copper connection to your house, full stop.
      How is this in any way problematic?
      I do have a slight problem with “up to” $5,000 because some installations would likely cost considerably more than this, though they would likely be ruled out under the “where practicable” clause in the coalition document. As an average though, it is probably not too bad since for the short (cheap) runs, copper is probably adequate for many people today. It is the long (expensive) runs that will be the first to want a fibre extension and this would drive up the average cost.
      Remember your geometry too, assuming a perfect circle radiating cable to all, pi r squared means 4 times as many people will live at more than half than average cable run than less than half. Given that cable runs don’t work like spokes in a wheel, that number will be higher.

    • I heard that the Coalition’s plan would cost up to 3 gajillion dollars. But would most likely come in under. We all know what a politician’s promise is worth though.

  4. At what point does either party question the minister in charge of such a important sector(not saying its more important the health etc)?
    Does it not prove suitable to place someone more specialised with the industry as the minister and not “the bloke whose been there the longest”?
    I understand not all the public understand completely the complexity of the NBN rollout but for those of us who do, why isnt it possible to have someone informed in the area debating these matters?

  5. I wish wish WISH Albanese knew his stuff enough to have called Turnbull out on that. Alberici did a good job asking him what she was supposed to to give Albanese his in….and he failed miserably.

    Turnbull gets the mainstream lie and while Renai and similar tech journalists attempt to correct it, the damage is done to the layman. Mr Albanese needs to research his new portfolio better. He should have at LEAST known 1Gbps AVC was $150 if nothing else about contention ratios, when asked directly.

    This is a classic case of Turnbull using his superior knowledge to misleadingly and deliberately turn a benefit of the NBN into a perceived failing.

    Thanks for writing this up Renai. At least someone wants to hold him accountable for what appears to be an increasingly misleading debate from BOTH sides. And Mr Albanese needs to do his homework- how can you sell your policy if you don’t fully understand it??

    • “Mr Albanese needs to research his new portfolio better.”

      He doesn’t have enough time to do that. It’s better if Albanese just avoided taking Turnbull on in a debate. Albanese should just hope that he isn’t invited to a National Press Club debate which would last for an hour.

    • +1. I was worried as soon as Albanese was made Minister for comms that this is the type of performance we could expect to see. In fairness to Albanese he’s had fairly limited to time to educate himself on the complicated nuances of the NBN (and he has multiple portfolios and an election campaign to worry about) but it’s still disappointing that he’s not able to call out Turnbull’s cleverly crafted lies.

    • I’d don’t think things like discussing CVC price are going to be on a wavelength that normal people can understand.

      All Albo had to do was to say “NBN Co are expecting a wholesale cost of around $150…I laugh in the face of your nobody-understands-you jargon MT ha ha ha!!”

    • Based on that Renai will have out grown FTTN by 2015 unless he lucks out and gets 100Mb, in which case HFC will also be a problem by 2016. By 2022 he will have to be forking out $20K for his 1Gb CVC

        • Year 1: Current bandwidth 26Mbps

          Result: Your predicted future bandwidth for the year 2014 is 39Mbps – This is based on starting date of 26Mbps

        • Plug your current bandwidth into the bandwidth growth calculated for expected bandwidth growth.
          I tried it on bandwidths from the early 90s that I had (taking 2013 as 1993, it’s pretty much been 50% a year growth for the last 30 years) and it was pretty much spot on.

          • thats interesting, even considering thats theoretical best growth…. 10y from 56K gets 3Mbit, ten from that gives 170-180odd mbit, and ten again somewhere just over 10gbps. there are caveats, being theoretical best it doesnt account for market strangulation effects *coughtelstracough*. can still be useful tho, im thinking. not everyone has bailed on 56K at the same time, and wont bail on dsl level speeds for some time yet, so theres overlap if you are an early adopter or mass accepter. but most users have clearly moved on to at least the early second decadal band, and once 1gbps gets released people will be able to order speeds better than 3mbit+10y.

            i get the feeling it probably indicates had our market developed a different way speeds today would be higher, tho. given how long 3mbits has been available for (its no doubt predated, but Telstra uncapping DSL to 8mbit in 2006 is probably fair) the calc seems to suggest a bandwidth of 51mbit would be appropriate “today” – 7 years from a 3mbit start. – and we know from the national average speeds there arent many people that fast.

            in any case plugging my current line speed (50 mbps, which isnt hit too often admittedly) suggests i should get the same utility out of my service at 2020 with an 850odd mbit line speed….. far above what the LNP offering promises to be on that date. even halving current speed to a still comfy 25mbit suggests similar utility at short on 450Mbit at 2020. still not close enough…. if you are comfortable with 4mbits now, you should be right with 68 mbits in 2020? thats in the ballpark of their promised mark?

            if the utility of the connection stays proportional, and bandwidth actually increases over those years as per the rate in the calc, thats basically saying my not-too-long junked wireless connection at 4mbit (nevermind it was only 0.01% of the time actually MAKING that speed) is roughly what i might expect the quality of connection under FTTN in 2020? the phrase ‘couldnt pull the skin off a custard’ comes to mind.

          • I usually start my figuring from 56k in 1999, and double every 2 years. That gets us to around the 12 Mbps speed they seem to think is the minimum now.

            Over that time, I’ve gone from 56k dialup, through 256k ADSL, and ultimately on to the 24 Mbps ADSL2 I currently ‘enjoy’ – actually 6 Mbps, on a good day. Point being, if you double the need for speed every 2 years, it really does follow Moore’s Law quite consistently, and at more than one point in the past (usually where theres a change of tech such as from dialup to ADSL) there has been discussion on either that speed not being needed, or that we wont need any more.

            So far every prediction of reaching a limit has been seen to be wrong. Totally, 100% wrong.

            I just cant see why that will change. It doesnt matter that people cant see a need, they couldnt see a need for cable speeds in 1999 either, yet it was there as an option. The speeds offered at the time (cap of 8Mbps from memory) had little to no residential use at that time, yet look back now and the speeds offered then couldnt survive today.

            The one thing I keep reminding myself is that even with the industry dragging their feet on rollouts, we’ve managed to roughly double our needs every 2 years. Now we’re at another crossroads, jumping from ADSL2 to FttX, I’d expect the industry (wholesale and retail) to at least look at those past lessons, and back the one that’s most cost effective to them.

            At which point you have to wonder why they would be happy with a model you KNOW will need more investment inside of 8 years.

  6. Renai, just a few quick notes with your calculations of the cost of a 1Gbps service:

    Firstly, the NBN Co wholesale costs are ex GST so the retail costs would be 10% higher than you calculate when GST is added in.

    Secondly, NBN Co’s report to the parliamentary joint committee (http://www.nbnco.com.au/assets/media-releases/2013/report-to-parliamentary-joint-committee.pdf) had a price graph for a 1Gbps service with 1000:1 contention (see page 14). Whilst I don’t know what the point is of having a service with a monthly quota that can be used in 40 minutes NBN Co seem to view that as an appropriate contention ratio for a 1Gbps service.

      • Sorry Renai, I didn’t see your reply until now.

        The graph on page 14 shows the price projection for a 1Gbps plan with 1Mbps of CVC (starting price of $170/month= $150 for a 1Gbps AVC + $20 for 1Mbps of CVC) which NBN Co equate to a 300GB quota (1Mbps * 60secs/min * 60min/hr * 24hr/day * 28day/month * 0.125bits/byte = 302.4GB)

        A 1Gbps plan with 1Mbps of CVC (300GB quota) is 1000:1 contention.

    • @Tallweirdo

      While I certainly appreciate we’re trying to keep things accurate here:

      1- 10% GST being missed on a $1200 charge is practically insignificant when comparing the $1200 of contended 1Gbps compared with Malcolms $20k uncontended. It is about 1%

      2- While I certainly believe that a 1:1000 contention on consumer models is excessive and unrealistic, I don’t think that was the point of NBNCo-s graph. It was a very poor comparison to make but it was essentially keeping the X and Y static by maintaining the z axis or outside variable of quota. I don’t think it was meant to be at all indicative of retail pricing. It was more to compare the lower tiers.

      • Yes 1000:1 does seem excessive. But you cannot take a contention ratio for a 12Mb service and say a 1Gb service requires the same. Over a large customer base speed becomes less the issue and the GB a month used is the dominating factor. In which case a 12Mb and 1Gb plan of say 200GB a month quota require about the same CVC.

  7. Found this comment on Facebook regarding pricing for a 1Gbps service.

    Australia wants a Fibre NBN: “As for the price of a 1000Mbps service for households, the AVC cost is $150 per month. This is equivalent to line rental.

    Then you have CVC. That is the ‘usage toll’ cost of data from you to your RSP, and also from your RSP back to you. Now if you need a Terabyte (1000 Gigabytes) of data per month, assuming 30 days in a month you need about 3 megabits of CVC. A megabit per month costs an RSP $20. Lets be safe and assume 4 Megabits so $80 per month.

    So the basic network cost of a gigabit plan with the capability of 1TB per month of data is around $230/month.

    Next you need to add data costs. Generally, 1 TB plans are available from some RSP’s for around the $100-$120 mark, including the AVC costs of that plan (100Mbps is $38/mo). So at worst, data cost is $80.

    We are now up to $150 + $80 + $80. That’s $310/mo.

    Next we add other costs like IP addresses, support staff and profit. I’d expect this to come into a margin of around 35-45% depending on RSP. Lets say $40[sic], so 1.4* 310 = ~$435/mo

    Suffice to say that no household will really be able to afford such a connection. But businesses pay much more for quite a bit less now. Internode/iiNet 10/10 connections (where available) are $750/month for 500GB data, or unlimited 10/10 for $1000/mo. Some massive benefits for businesses in terms of reduction in costs and much faster speeds.”

    With 4 Mbps of CVC in the example given, does that mean the contention ratio is 1:250? Is that an unrealistic contention ratio? I’m also not particularly sure why they related the data usage (1000 TB) to the CVC.

    I found some more information here:
    http://www.tektel.com.au/NBN-pricing-explained.shtml

    A couple things I’d like an explanation, or confirmation, for, though.
    Let’s say you have budget ISP A servicing a customer in an area, and premium ISP B servicing another customer in the same area, the former would purchase 4 Mbps and the latter would purchase 40 Mbps for each customer for a total of 44 Mbps reserved on the “link”?
    With the NBN GPON network, there are 32 fibres split off one link, right? What is the total capacity of the link? Let’s say that there were 2 businesses in one area that needed a 1:1 1Gbps service. 2 Gbps dedicated. Would the capacity be enough?
    Let’s say then that there were 20. So that’s 20Gbps dedicated. I’m assuming that the guaranteed bandwidth would exceed the limit of the link. What would the RSPs, or NBN Co, do in that instance?

    Would it be fair and accurate to say that CVC relates to minimum speeds and speeds at peak usage?

    • Okay, I understand now how (though not really why) they related CVC to data allowance. It’s just a unit conversion.
      1000 GB/mo
      8000 Gb/mo
      8192000 Mb/mo
      273066.6667 Mb/day
      11377.77778 Mb/hour
      189.6296296 Mb/min
      3.160493827 Mb/sec

        • Yep, peaks a killer.
          BUT, since NBNCo has speeds that allow RSPs to sell really fast big quota connections to latger companies, hopefully there will be more usage of that CVC during the day too.

        • Yeah, well, as I said, I understood how they calculated it, but didn’t really understand why.

          I guess that, in a way, it’s a guarantee that you’ll be able to use your entire quota each month, if you so wish.

    • “Suffice to say that no household will really be able to afford such a connection. ”

      BS

      $400/month is not an excessively large cost – it’s about $13/day.

      Of course, one would need to find a compelling reason to need pay this over a cheaper alternatives – but these aren’t hard to come by. Working from home is an obvious one, as is entertainment (especially for wealthy or multi-tenant properties e.g. share house) – and of course medical costs are obvious.

      Maybe it’s just who I am, because work from home managing remote teams and clients, manage a 7 bedroom share house, and work with a nursing company that provides 24×7 care… but seeing opportunities for 1Gbps speeds in residential use isn’t hard at all.

      • The commenter (the person who runs the Facebook page “Australia wants a Fibre NBN”) was just explaining to a lady how much a 1Gbps connection would cost, as she had already called bullshit on Malcolm’s $20k figure, but she didn’t have the facts, so she wanted to know how much it might actually cost.

        I think it is true that the $430 charge isn’t really affordable for a “household”, you know, a family, or maybe a group of friends living together. I can’t think of any other monthly charges that are in that range. In terms of adult wages here in Perth, that’s something like 21 or more hours of work. It’d be worse with the lower wages elsewhere. If you split it among say 3 people, it’s probably not a big deal I guess, a day’s wages each. He wasn’t talking about edge cases like home businesses, and I don’t know if you’d call a share-house a household, but I guess you could. (I, for one, can’t wait until everyone starts working from home.)

        Ultimately though, he just wanted to answer the lady’s question in the context in which she asked it, and I think that’s important.

        What upsets me is that there are so many people (on Malcolm Turnbull’s Facebook page) who simply accepted the ludicrous $20k figure as being true, instead of more people like that lady who heard the $20k figure and went “uh, hang on…”. I mean come on, alarm bells don’t go off in your head when you hear that figure?

        Yeah, I know. I need to get off Facebook.

        • “I think it is true that the $430 charge isn’t really affordable for a “household”, you know, a family, or maybe a group of friends living together. I can’t think of any other monthly charges that are in that range.”

          Really?

          How about rent? That can easily range between $1200 to $2600/month.
          How about electricity? I’m looking at paying around $300+ a month for that.
          How about owning and operating multiple vehicles?
          How about regularly going out?
          How about owning unnecessary office space… that can be stupidly expensive.

          “In terms of adult wages here in Perth, that’s something like 21 or more hours of work. ”

          So no adult in perth gets paid more than ~$20/hr ever? What crappy place to be!

          ——

          Seriously people please learn to *think*. There’s a huge difference between being able for afford and being able to find the value in it, there’s a huge difference between some households and whatever the defined ‘average’ household is, and there’s a huge difference between the value that can be found today in high-speed communications and the value that can be found tomorrow, when they are available and open to markets of scale.

          • I’m just talking my impressions of the averages here, and after tax. Some make even less, some are single-income households, some make more, some are multi-income households.

            Look, everything you say is correct. I’m not saying you’re wrong. You really don’t need to take that attitude with me. I had certainly overlooked those charges (sort of), but you usually consider the affordability of something AFTER you take care of rent, utilities, transport, food, basic internet and phone – after you take care of the necessities. What income you have left is your “disposable income” (not by the official economic definition, mind), and that’s where you have to make choices – the more expensive the choice (e.g. a $430 1Gbps connection), the more of the other choices you have to do without. Affordability is not divorced of value, affordability is not just ‘can we afford to have X’ but also ‘can we afford not to have X’, and what determines the answer is relative value. Compared to what households are /typically/ paying for their internet today, $430 represents a significant increase. Would people want it? Probably, yes. Would they actually get it? Only if they could ‘afford’ it.

            I’m not at all arguing against 1Gbps (or better) services. I welcome it. I know we’ll need it in the future – and many people will benefit from it today. The context is different though. As I said, the commenter was answering the lady’s question in the context in which she asked it (essentially, how much would it cost a household today?). You have changed the context to include other types of residences, and include future use. What you say is right, but what I’m saying is not wrong either.

            Personally I see a 250Mbps service as likely being a lot more attractive (today); doing the above calculation, I’ll say somewhere around 1.4*(70+60+60) ~ $270; sure it’s 60% of the calculated 1Gbps price, for only 1/4 the speed, but at the same time… you’d have to regularly max out your pipe to get value out of it (and if you’re doing that, you probably need a bigger quota).

            Off-topic, with your electricity bill, that means you’re paying more than $10k every 3 years? Are you able to get rooftop solar? I’m living in a family of 7 and our last monthly bill was $280, so not that far off yours, and we’ve just started looking into it.

        • Your $430 is for a “best effort” connection, though.

          NBNCo will be offering high priority traffic, enterprise grade, etc. from 2014. But the costs are astronomical, same as private sector now.

          • @Michael

            I disagree entirely. $25K for a uncontended 1Gbps pipe is chicken feed. Enterprises pay hundreds of thousands a month for that now.

            Even $2500 for a 1:10 1Gbps pipe is brilliant value for a medium business wanting high bandwidth but not as a guaranteed pipe for interconnecting offices- just for employee internet requirements etc.

            10/10Mbps contended Ethernet from Telstra is $3500 a month.

          • No surprise Michael is talking out of his a#&* again here though, is it?

            Seriously Michael, spend 15mins making a few phone calls to find out the cost of connecting and operating a dedicated fibre line to an ISP backbone in the CBD of your choice. Then compare with the costs of an NBN line. Large enterprise will still require these dedicated connections, but for plenty of smaller operations NBN fibre will be game changing.

            While unfortunately such research won’t play to your biassed position, at least if you report your findings honestly you will have the satisfaction of knowing thousands of people who read it all over the country will be utterly gobsmacked.

          • Oh, it is absolutely cheaper than many would pay for such a link currently. It’s still expensive, though, my point was that it’s not fair to compare to the strictly consumer grade service.

            btw, links of such a high bandwidth and low contention are almost always quoted. There is no set price from current operators. NBN changes that of course because it’s all the same infrastructure and set up the same way. But currently: I have seen 1gbps corporate fibre services for ~$15k in the CBD and 100M/100M internet for ~$1k also in the CBD. I’ve also seen 1M/1M ethernet over copper quoted at $10k in a rural area (!). It’s all over the place.

          • @Michael

            $15K in city for 1Gbps? Unlimited data? Uncontended link? Per month?

            I find that VERY hard to believe.

            That’s what the $20K on the NBN would be equivalent to- an uncontended, quotaless (because it has 1:1 CVC there’s no reason it shouldn’t have unlimited data) link per month.

          • I’m not sure where you got your pricing from but we have dark fibre and it costs no where near the prices you’ve quoted, even our 100Mbit wan links as a yearly fee isn’t what you quoted as a monthly fee.

            On the darkfibre ~60km run we can run anywhere from 1-10Gbit depending on CWDM/EWDM, they don’t really care it’s the fibre you consume and the distance it’s run.

          • @Jason

            Ok. Cool. I don’t obviously expect you to provide a bill to prove it :P

            But can you give an idea of what pipe you have and how much it costs? I’m not expecting SLA details, just the speed and symmetry of the line and monthly cost?

          • And that is what will push up ARPU and revenue/profit allowing reductions in wholesale pricing

          • No one claimed otherwise.

            But looking at it in the context of this question,
            “For the average household?” asked Alberici.

            As Lionel pointed out,
            Over a large customer base speed becomes less the issue and the GB a month used is the dominating factor. In which case a 12Mb and 1Gb plan of say 200GB a month quota require about the same CVC.

            $430 might therefore be a price that average households could reasonably be expected to pay for a 1TB 1Gbps connection.

            Whether an average household would need a 1Gbps connection today, and whether 1TB would even be a reasonably sufficient quota for a 1Gbps connection, those are different questions.

  8. as far as the biggest off goes i was surprised this one wasnt quoted:

    ” “Nobody has been able to say our assumptions are unreasonable or that given those assumptions that you would not end up with a $94 billion figure,” Turnbull said.

    “You’d think if they were wrong in more than four months [since the Coalition raised the number], the Government or NBN Co or someone would have said, ‘No Malcolm, you’ve got that number wrong’. No one has said that.””

    umm is it just me or has that 90bn figure been done multiple times, and MT has conveniently ignored, ‘didnt remember’ or ‘hasnt seen that’ because its – lets face it – convenient to him? (fairly sure Renais pointed to the fact that figure only holds if every worst possible thing happened to the build, and i suspect steveJ and Nick Ross might have as well, but i dont remember off the top of my head. if im right on all counts thats at least three times it has been said. Dont remember if Conroy or Quigley direct refuted it – i think Quigley might have in an estimates hearing?).

    with the way steveJ got treated the other day (“noone reads your shit”, basically) i get the impression that MT/staffers are working from a closed universe so they can evade discussions and factchecks by begging ignorance.

    and that gets right up my nose – presenting it as a fait accompli: “well, noones pulled me up on it therefore i must be right!”. apparent absence of criticism isnt much of a marker if its clear you arent bothered looking for/at any…..

    • I Believe MQ did a complete destruct on that Spin Figure in the committee, Maybe M.T missed coz he was daydreaming or maybe had his fingers in his ears, so even bigger lie, I was disappointed Albanese failed to challenge that claim and let it pass without a whimper

  9. 1 gig plans will be controlled by download limits. If you had a 2 TB download limit the 1 GB per sec plan would only be active 0.8% of the time. I suspect a contention rate of 50 to 1 would give adequate service, especially as most people only use a fraction of their quota.

    Thats $40 a month ($2000/50) for cvc costs (wholesale)

  10. thats the problem with pollies the average joe blow would believe his lie of 20000 grand.i couldnt believe it wen i heard him say that.all turnball is doing is going for votes saying his is cheaper and quicker TO build.may be a few billion cheaper but its inferior teck.i think the residential cost of 1gb will be about $120 to $150 depending on isp.i hate politics just hope ruddy wins only cause of the NBN.

  11. What I want to know is how true is Turnbull’s claim of NBN’s projected wholesale price increases and increased revenue has been factored in to the total cost of the rollout? Has NBN reduced the total material and labour costs based on expected revenues? If Turnbull is correct it is a worry that in the not to distant future wholesale price increases of NBN connections will increase significantly compared to where they are today.
    One thing that’s clearly evident is that there is a problem what the contractors are being paid. As it is I don’t think the project can stay on track with this problem.

    • @Rob5089

      If you look at the NBNCo. Corporate Plan, you’ll see the Average Revenue Per User increases 3-fold over 10 years. But what Turnbull DOESN’T say is that the ARPU is NOT rising because of an increase in prices. It is rising because NBNCo. project multiple wholesale products for a portion of the market and high-end wholesale products being taken up.

      For example, if the 2 million SMEs in Australia took up 100Mbps over the next 10 years, that would be $38 per user straight away, plus all the CVC associated with them (at an average 1:10 contention, that’s ~$150/user once CVC price reductions are taken into account). That’s $188 per user for 2 million users. Even if the 9 million households they expect to take up the NBN to 2023 only took up an average of 25Mbps, at 1:30 contention, that’s 9 million at $35/user + 2 million at $188 per user. That’s $63 ARPU already. That’s 3 times HIGHER that it is now and NONE of that is unreasonable at all. AND, most importantly, NO ONE has seen an increase in cost of their broadband plan. In fact, retail prices will have DROPPED thanks to CVC prices dropping over that time.

      Add in Enterprise 1Gbps connections, multiple business connections per NTD, work/education/health/security products for consumers…..you can see how $122/user could easily be reached in 10 years and STILL user would be paying $50 or less a month for their 25Mbps.

      That line from Turnbull about ARPU is a scare tactic. It assumes voters don’t understand business….which is a reasonable assumption of many. But in reality, NBNCo. expects everything from BB, security, [hone, health, education, business and PayTV and more to be on the NBN. It will, after all, be THE only fixed-line network in 93% of Australia. And it will be FAR more flexible and offer FAR more products than Telstra ever could on copper. ARPU WILL rise. And retail BB prices WILL drop. It is perfectly reasonable. Turnbull would just like you to believe it isn’t.

      • Thanks for clearing that up seven tech. The technical argument is easy for myself to understand being a techie but talking about the business side of things it’s harder to decipher the truth. It makes it really hard when Malcolm is really well spoken in this area and Albanese and even his predecessor Conroy does not reply back with the same finesse and detail.

        • @Rob5089

          It’s one of Turnbull’s biggest strengths- he knows and understands what he is talking about. Most politicians wouldn’t know that level of detail about a specific policy. That’s where Albanese and even Conroy to an extent fell down.

          But I think UNLIKE Albanese, Conroy would’ve wiped the floor with Turnbull over the ARPU issue. It’s actually very easy to counter-act with an analogy- On 2 different days, McDonald’s Average Revenue per Buyer is 20% different. That is, the average amount a person spent at that McDonald’s was say $12 on Monday and $13.20 on Tuesday. NOBODY would believe that was because prices rose by 20% in one day. Stick that same argument over 10 years though and without thinking, 90% of people would believe it’s because prices rose. When you actually analyses NBNCo’s plan though, prices DROP while more services that are already being used OFF the NBN, move ONTO the NBN and make up for the dropping average wholesale price to increase the ARPU.

    • Rob
      Two things
      1) NBNCo has stated that as revenue and profitability increase wholesale charges will be REDUCED, many suggest that rather than just reduce charges, more is spent on expanding the footprint and upgrading
      2)Contractor payments. That is not as simple and clear cut as it appears. The issue has much to do with the patchy stop start nature of the workload up to this time impacting the contractors who then have to squeeze the subbies beneath them. A relatively steady workload would have worked for all parties including the subbies who could have been paid more. Remember it is not NBNCo employing and paying the contractors and subbies. On record – NBNCo being charged $106/Metre for Horizontal drilling by the prime contractor and paying that price, the drilling contractor being paid more like $40/meter or even less

    • Now you wouldn’t be trying to derail the commentary about the article would you?

      Both sides of politics need to pull up their socks, but then again, they are pollies, paid to stretch the truth so thin that the obvious underlying lies are apparent and easily seen under that thin veneer of truth.

  12. I think that the estimates here of the cost of a 1Gbps service are quite inflated for two reasons:

    1. The CVC required applies to an entire Point of Interconnect (PoI). With large PoIs the contention ratios can be much higher as the peaks from different types of customers (business use in the day versus home use at night, etc) will balance each other out more effectively. This is a bit different from your average suburban telephone exchange (which would generally have a much smaller number of customers and most of them will probably be residential and all accessing their services at once).
    2. Even if customers do have a 1Gbps connection the odds of them finding servers on the internet that can deliver them content at that speed are not great. This will change over time (particularly as new products are developed to take advantage of the NBN), but by then CVC prices will probably have come down anyway. The odds of even a handful of customers on a single PoI finding a server that can deliver content at 1Gbps all at once is quite low.

    For these reasons I think that the contention ratios can probably be much, much higher than people are expecting.

  13. IMO the focus on the 1Gb at this time, especially for domestic use is pointless, in fact IMO counter productive as the voter sees it as “Gold Plating” even though just demonstrating the flexibility inherent with FTTP .
    The 1Gb was always in the NBN plan but for business along with the 250 and 500 So not in response to GOOGLE even though the media waffle over Google’s superior 1Gb meant highlighting the product and releasing it a year or so earlier than planned, however a business that could use that could be from a residential dwelling or premises within a residential area not necessarily within a business or industrial precinct.

    The absolutely key differential has not been discussed and that is not a Data stream Factor, but rather a transmission factor and as such has not truly been appreciated.
    That is that with FTTN the 2.5 Gb SPECTRUM available (which can include a spectrum (wavelength) band dedicated to multicast at a reduced cost) will be decoded into discrete streams per user for Electrical transmission on pairs or encoding back to light for FTTP . Certainly some of that could be multicast, but splitting 25Mb or oven 100Mb for a multi user premises is limiting .

    However with GPON FTTP that whole 2.5Gb (or 10, or 40, or 100) is split to each subscriber and the NTU decodes their service, that 4 port NTU can decode 4 discrete separate services.
    Remember 10Gpon now available 40GPON soon and 100GPON by the time it is built.

    This may appear inconsequential at this time but so many possibilities are made possible apart from just multicast Video including high quality 4K not just low quality compressed versions, such as public service (government services), community services (within the Fan) etc etc.

    It is the capability for future evolution of services that is the differentiating factor and I am disappointed the Labor politicians fail to comprehend and sell the National Information Transmission Network, of which “Broadband” is a large COMPONENT.
    On the basis of a transmission Network the LBN starts to look shaky.

    P.S my background as you may guess is Transmission, not IT

  14. Dear Renai,

    Don’t be so quick to crucify Turnbull on this one.

    There are other issues here that the discussion format didn’t permit him to get into. Here’s just one of them:

    To enter the market – to hang your shingle out as an RSP before you sign anyone up – you need to buy enough CVC ‘size’ at each point of interconnect (POI) in order to deliver an acceptable service.

    There’s an obvious point here – to allow a customer to burst to their full line speed, the CVC has to be at least the same size as that customer line speed.

    Lets put some numbers on this – follow along with me, the maths isn’t hard.

    To enter the 100 megabit customer circuit market, you need at least 100 megabits of CVC purchased at each POI. So thats 100 x $20 = $2000 per month (per POI) x 121 POI’s = $242,000 per month nationally (or $2.9m per year). Thats the minimum cost to be in the market at all.

    To enter the 1000 megabit (gigabit port speed) market, nationally – to offer gigabit line speeds to customers, regardless of contention ratio, you need to buy at least 1 gigabit in every POI in the country (or else you’re selling a line speed you can’t deliver to your customer, ever, regardless of contention ratio).

    Lets do the maths:

    1000 x $20 = $20,000 per month (per POI) x 121 POI’s = $2,420,000 per month nationally (or $29m per year).

    An RSP who wants to offer national 1Gb/s customer services has to spend $29m per year buying unused air from NBNCo in terms of CVC charges, and has to soak that (massive) cost for several years, while the customer base/takeup rises to use the pipe size to a reasonable utilisation level.

    How many RSPs do you think are going to be prepared to pay that extra ($29m vs $2.9m per year) to hang out their shingle in the 1Gb/s customer port market?

    The fault here is in the NBNCo charging model. Not *just* the $20/Mbit CVC (which is 10x higher than it should be, against rational wholesale backhaul costs in high capacity fibre markets). But also that the capacity has to be statically purchased in this manner by RSPs.

    The fix is remarkably simple: It’d just need NBNCo to move to charging CVC’s using 95th percentile billing:

    http://en.wikipedia.org/wiki/Burstable_billing

    This is the standard billing mechanism in fibre-dense trunking environments of this sort, around the world.

    Except for NBNCo.

    Unless NBNCo makes this change, I contend (cough) that almost no RSP will offer 1Gb/s port speeds to customers for the next several years at least, and definitely not as a nationally consistent offering (maybe in one or two business-dense POI’s in major cities only).

    Naturally, the other way to massively increase the market participation by RSPs would be to collapse the 121 POI’s back to the original 7 pairs of dual-redundant access points that was in the original technical and business model as well (which is easy because those sites still exist and they’re all going to get built).

    Anyway – the bottom line, Renai, is that the CVC charge of $20,000 per Gigabit is, indeed, a massive dis-incentive to gigabit rate customer services takeup in the current NBNCo charging model for the reasons noted above.

    What you’ll see if this isn’t changed is claims that ‘nobody wants gigabit services’ on the basis that no RSP offers them – but really, its the cost model thats the broken thing here, not lack of demand.

    The demand will be there – just follow the Cisco VNI graphs in their inexorable rise over the next 10 years. But in Australia the NBNCo cost model will be suppressing that demand – and hence suppressing innovation and customer benefit – as it stands today.

    • > Anyway – the bottom line, Renai, is that the CVC charge of $20,000 per Gigabit is, indeed, a massive dis-incentive to gigabit rate customer services takeup in the current NBNCo charging model for the reasons noted above.

      Yep, NBN Co needs to have a legal obligation to reduce wholesale prices if there’s even a slightest hint of market demand rising up, other than for purposes of cost recovery.

      > How many RSPs do you think are going to be prepared to pay that extra ($29m vs $2.9m per year) to hang out their shingle in the 1Gb/s customer port market?

      I wouldn’t really call it ‘extra’, especially not in the long term, but it’s an up-front challenge, which had a pretty massive impact on Internode in particular even though it’s not really there yet.

      > The fix is remarkably simple: It’d just need NBNCo to move to charging CVC’s using 95th percentile billing:

      It’s not a permanent fix by any means if NBN Co has no legal compulsion to reduce wholesale prices faster than CPI-1.5%.

      Let’s face it, Malcolm’s claim of $20,000 a month implies someone downloading (and uploading a fair amount too) well in excess of 300 TB a month, which is near enough 10,000 times the current average. It turns out that the current fixed broadband average comes up to about $2 in CVC charges a month.

      I don’t think many people at all would mind if Internode/iiNet stick to 250 Mbps maximum for the next few years until download amounts come up and NBN Co reduces CVC pricing.

      http://i.imgur.com/kcFUkC9.png

      > maybe in one or two business-dense POI’s in major cities only

      My prayers of that service working out with a high contention ratio are with you.

    • “Don’t be so quick to crucify Turnbull on this one.”

      Why? He’s clearly lying.

      “To enter the 100 megabit customer circuit market, you need at least 100 megabits of CVC purchased at each POI. So thats 100 x $20 = $2000 per month (per POI) x 121 POI’s = $242,000 per month nationally (or $2.9m per year). Thats the minimum cost to be in the market at all.”

      Whoa, now we’re talking about an *RSP’s* cost to provide 1Gbps services across a nation, how does that fit in with $20,000 being an absolute cost for any single residential user?

      “How many RSPs do you think are going to be prepared to pay that extra ($29m vs $2.9m per year) to hang out their shingle in the 1Gb/s customer port market?”

      Potentially quite a few. You for some reason ignore the three obvious points:

      1) There may be a sufficient market.
      2) One doesn’t have to go nationwide straight away.
      3) If a smaller RSP can’t deliver the service through their own contracts with NBNco, they may be able to use a wholesaler.

      Lastly of course, it’s irrelevant how many offer this service. With NBNco and RSP *can* offer this service… which means at worst the threat of competition should keep the big players prices more reasonable. And of course, having the infrastructure in the ground (even at $20k a customer a month) is a lot lower entry bar than physically rolling out cable.

      You supposedly know more about this industry than me, please show it.

    • Simon,

      While I appreciate your position on this CVC issue, it does NOT give Turnbull the right to say, SPECIFICALLY when asked, that ordinary households would pay $20 000 for a 1Gbps pipe.

      You’ve said yourself CVC is an issue for RSPs because of the sheer scale you need to have for profit making. I understand that. I don’t necessarily agree on all points, but I do on some. But that doesn’t change the fact that there is no way in hell that a 1Gbps pipe provided on a CONSUMER basis would be $20K for every household. A CSA has to have 1Gbps before you can give 1Gbps AVC…..but that does NOT mean you can’t use that 1Gbps for contended CVC on lower tiers. Suggesting, as you appear to have, that you NEED a 1:1 contention on every POI to provide 1Gbps is misleading. Yes, you must provide 1Gbps CVC….but what RSP is going to provide ONLY 1Gbps and ONLY to a POI they haven’t yet supplied for??

      It will be a difficult transition for smaller RSPs to make and I would not expect them to offer 1Gbps plans in the near future. But Telstra? Iinet? Optus? Why not? If businesses are willing to pay 2 or 3 or 4 thousands a month for a reasonably uncontended 1Gbps GPON service on a CSA that already has a few thousand customers on it….why wouldn’t they?? It’s profit.

      I agree CVC is an issue for transition. But I DON’T believe that gives Turnbull the excuse to outright lie in suggesting 1Gbps for CONSUMERS would be $20K- no RSP in their right mind would think they would get any takers for that. And I also don’t believe CVC will be an issue as the NBN gets alot more scale.

      • > If businesses are willing to pay 2 or 3 or 4 thousands a month for a reasonably uncontended 1Gbps GPON service on a CSA that already has a few thousand customers on it….why wouldn’t they?? It’s profit.

        And I think there’s a real possibility that that business is likely to want that 1 Gbps for off-site backups, running at 2am in the morning.

        You know what I think might be worth it, on the RSP side?

        On-demand pricing. Just get it done. 1 GB at 7pm, let’s face it, should cost more than at 3am.

        If nothing else, on-demand pricing introduced at the top end would manage this whole brou-ha-ha about:

        > The demand will be there – just follow the Cisco VNI graphs in their inexorable rise over the next 10 years. But in Australia the NBNCo cost model will be suppressing that demand – and hence suppressing innovation and customer benefit – as it stands today.

        Yes, it’s an artificial scarcity designed to extract money. Deal with it. Is it of any consequence to your average user even with some years of exponential growth under the belt, if the current fixed broadband average comes up to about $2 in CVC charges a month? No.

        I hate to sound like everyone, not least of all Malcolm Turnbull, but what innovation requires 1 Gbps, completely uncontented, for the vast majority of the population within the next decade or even three? I think that even with Ookla and ABS and Cisco showing exponential growth, the CVC model isn’t going to be broken broken with no way to work around it and it being any sliver of a break of innovation.

        And not least of all, it pays for the damn thing. You want 100GPON? Well, you better pay your CVC charges to pay off at least a substantial fraction of the current GPON first.

    • Simon, assuming you don’t offer 1Gb right off the bat, what portion of market share would you need before your CVC would be high enough through your customer base that you could start offering them?
      This POI ACCC decisions seems to have hurt a lot of players to appease the few like Pipe, etc.

      • Actually, I guess another way around it is to change the pricing model to a data usage model. Maybe have peak/off peak, let the ISP decide if that changes his plans or he just goes flat and averages costs over the peak/off peak.

        Are they simply using a speed model because the rest of the industry does, eg, backhaul, overseas capacity, etc?

        • My apologises, it looks like, if it gets accepted, the post will actually be below. It’s awaiting moderation.

          Probably because it’s quite long and I make a few pointed (but in good faith) remarks to Simon. Hopefully Renai feels it is appropriate. He’s been very strict on his moderation of late. If he doesn’t post it, I’ll post a brief explanation of how I arrived at the 5% figure for you.

      • Lionel
        “This POI ACCC decisions seems to have hurt a lot of players to appease the few like Pipe, etc.”

        Very true, the result is actually anti competition and innovation.

        This concerns me that if the ACCC can unintentionally get it so wrong, what hope do we have of the Productivity Commission or who ever will do the Coalitions CBA actually being able to analyse all the future benefits of the Current Multifaceted National Communications network as being built , especially as compared with the simplistic Data model of the FTTN model.
        Sure some benefits can be approximated by software emulation. Windows on Linux or Apple anyone

    • Thanks for the excellent explanation.

      Is it technically and legislatively feasible to actually reduce the POIs back down to 14? I’m not doubting you here, I’m genuinely curious. How would it happen?

      I’ll note though, that your comment didn’t really have anything to do with Turnbull’s lie; at best they shared a topic (CVC charges). I’ll forgive it as being a creative segue into your comment though.

    • Don’t be so quick to crucify Turnbull on this one.

      Renai has been very careful not to call Turnbull out on his misrepresentations, because he isn’t sure if Turnbull is being delibrately deceptive. The fact that he has “crucified” Turnbull here is an indication, to me at least, that Renai has considered this very seriously indeed.

      There’s an obvious point here – to allow a customer to burst to their full line speed, the CVC has to be at least the same size as that customer line speed.

      To enter the 1000 megabit (gigabit port speed) market, nationally – to offer gigabit line speeds to customers, regardless of contention ratio, you need to buy at least 1 gigabit in every POI in the country (or else you’re selling a line speed you can’t deliver to your customer, ever, regardless of contention ratio).

      This assumption is one I contest: that you must provide services to 100% of the country.

      You don’t. In fact, you can wait until you have at least 2.1 Gigabits CVC in a given PoI (that is 150% of burst capacity for a 1000/400 connection), which is, assuming a contention ratio of 140:1 on 100Mbps, 70:1 on 50Mbps, and 35:1 on a 25Mbps connection and 13:1 on a 12Mbps (I use these numbers for as they represent exactly 1Mbps per user, however they are pretty close to real world figures you would expect) that means you only need 2,100 customers per POI. Since there are, assuming even distribution about 82,600 possible households per PoI, this only represents 2.6% market share in that PoI.

      So Simon, you could even, for safety sake, double that, and say “unless we have 5% of the market share in your PoI you cannot buy a 1Gbps service from us”. This is a perfectly reasonable response to take, and one your customers will respect.

      There is another model, which I pointed out to you on Twitter, and this is the “We will only offer it if enough people ask for it”. If we assume you’re going to offer 250:1 contention, and charge them slightly above cost, that means you will only NEED 250 customers to sign up in a given PoI to make a modest profit. But that’s a considerable risk.

      Instead you’d probably want 500 in a given PoI to agree to sign up. That assumes a rather generous failure rate of 50% of people who said they’d sign up, but don’t end up doing it. With this model you would probably end up charging more than the former model (because in the former model you already have the requisite CVC in your “pool” from your other, lower tier customers), but you don’t have to worry about unpredictable demand, particularity if you’re upfront with how much it is going to cost. (And also, if you get enough customers in lower tiers you can always lower the price).

      With a contention ratio of 250:1, this results in $230/m per customer wholesale costs to NBNCo. That means, taking into other factors, you can probably charge between $300/m and $350/m and still turn a modest profit.

      The fault here is in the NBNCo charging model. Not *just* the $20/Mbit CVC (which is 10x higher than it should be, against rational wholesale backhaul costs in high capacity fibre markets). But also that the capacity has to be statically purchased in this manner by RSPs.

      While there are flaws with the NBNCo’s pricing model, this does not mean a RSP offering 1Gbps, even one as small as Internode, is some big mountain to climb. You just need to start thinking outside the box. The first thing I would consider Simon is quite simple:

      Be honest with your customers about what it costs you to provide a given tier of service.

      If you can’t afford to give them 1Gbps because you don’t have 5% market share in their PoI, tell them. If you can’t afford to give them 1Gbps because you don’t have 499 other customers who want it too, tell them.

      Customers in todays interconnected world a very resourceful. You might find they do all of your marketing for you, and get people to churn for you, such that you can meet the ratios you require in order to provide them the service they require.

      There is no point dumbing it down for them. You are not Telstra, or Optus, and you don’t have to be either. Be an ISP for the “geeks”. Sell Unlimited plans that have a set contention ratio instead of a set quota. Give customers usage heat maps so they may self shape to fit demand. Just remember: give a man a fish, and you’ll feed him for a day. Teach a man to fish and you’ll feed him for a lifetime. You don’t NEED to be the one who crunches all the numbers and works out how much CVC customers need. You can get the customers to do that themselves; if you give them the right tools.

      Unless NBNCo makes this change, I contend (cough) that almost no RSP will offer 1Gb/s port speeds to customers for the next several years at least, and definitely not as a nationally consistent offering (maybe in one or two business-dense POI’s in major cities only).

      This is because you are assuming that the current models of deliver will survive, or even have to survive. An open wholesale monopoly gives you room to experiment, so let us see some experimentation. I want to see some “Only if we have enough demand” Google Fibre like Broadband offerings. I want to see some “Pay per Gigabyte” offerings. I want to see some “You choose your CVC and we give you heat maps and other data to crunch so you can maximise your connection” offerings.

      I don’t want too see “100/40Mbps with 200GB of quota per month”. That is boring. We’ve seen it all before. We know hardly anyone actually uses their full quota, we know that your quota is in no way reflective of your back-haul, but with the exception of a few experiments with Exetel that didn’t pan out and TPG remodelling what it means to be “Unlimited”, we’re stuck in the same product offerings we were 10 years ago. The only difference is the numbers have gotten a little bigger.

      Naturally, the other way to massively increase the market participation by RSPs would be to collapse the 121 POI’s back to the original 7 pairs of dual-redundant access points that was in the original technical and business model as well (which is easy because those sites still exist and they’re all going to get built).

      Unfortunately that is also wishful thinking Simon.

      Anyway – the bottom line, Renai, is that the CVC charge of $20,000 per Gigabit is, indeed, a massive dis-incentive to gigabit rate customer services takeup in the current NBNCo charging model for the reasons noted above.

      This may seem harsh: but business’ need to innovate, or die. The only reason you can’t find creative ways to offer 1Gbps to as many people as possible is because you are either scared, or unwilling, too take risks. Now as a privately owned company, or at least you were until you sold out to iiBorg, you don’t have to sell out to shareholders who like nice, safe, investments.

      You have an opportunity to completely refine the way Internet is provided to Australians, and you’re complaining that profit wasn’t handed to you on a silver plater by the government? I expect better from you Simon. We’re more likely to see interesting innovations from John Linton if you really think this is a mountain you can’t climb.

      • @NightKhaos

        100% agree with everything you have said there. It is a consumer market and those that consume determine the consumption. Trying to fit them into in a neat little box and stuffing it out the way to pull money from time to time is not a way to happy customers, a profitable business or a good reputation.

        Simon has done some fantastic things for the ISP business by innovating and making money where Telstra made it almost impossible to do so. Is he truly now saying in a MORE open, MORE competitive and equal market on the NBN, these challenges are far too hard to overcome??

        People have campaigned for the NBN for years. Now that it’s here, because government has to pay and wants to make the money back, rather than just gifting it to the industry, suddenly the NBN isn’t the competitive platform they wanted? There’s no question that CVC is an ongoing challenge for RSPs to overcome. But with scale comes ease. And nobody could ever say Internode is a ‘small’ RSP. I did modelling 18 months ago that showed when an RSP has 10k customers per CSA, CVC makes up approx. 10% of the per user price. 10k users would be 1.2 million….or about 10% of the market. If I’m not mistaken, iiborg already has nearly 25%….and 10% is hardly onerous. It goes down to less than 7% at 20k.

      • > If you can’t afford to give them 1Gbps because you don’t have 5% market share in their PoI, tell them. If you can’t afford to give them 1Gbps because you don’t have 499 other customers who want it too, tell them.

        I’d argue it’s not even that complex.

        Once you’ve got a fair number of users on 100 Mbps and maybe some on 250 Mbps, you’ll very likely be close to having the CVC in a POI to sustain some small quantity of services at 1000 Mbps.

        If you want to offer 1 Gbps, you don’t need 499 other customers on 1 Gbps, the statistics might well work out to show that 499 customers on 100 Mbps might be far more than enough to provide that service already.

        • I was referring specifically to the “we’ll only provide it with 500+ users agreeing to pay for it” scenario that I suggested above with that half of the comment.

          If they have enough customers (over 2100 assuming the contention ratios I suggested) they do have enough CVC for some 1Gbps customers already.

  15. Arguments aside, Turnball appeared to present his arguments better than Albanese and Labor are setting themselves up to get stung when he said that it would be delivered on time on budget. I don’t think any large government project has been delivered ontime and onbudget

  16. Malcolm didn’t lie as a true 1Gbps 1:1 would cost that much but as you said it is unrealistic to believe this cost would be transferred to the customer. What annoys me most is the host didn’t allow proper explaining as to what these claims are based in. She consistently cut off both Turnbull and Albanese when they didn’t give a 1 word response which in turn would be misleading.

  17. The whole interview was horrible from the interviewer to the interviewees. I think Turnbull was somewhat clear “if you want to have a guaranteed one gig service” he was being a little slippery as this is rarely going to be a service anyone subscribes to. He deserved a little bit of criticism. I just wish these high standards applied to all NBN topics and issues on delimiter. This website is very frustrating to read as it overlooks some blatant deception by some and issues facing the NBN. I still thank you for your contribution as it is better than nothing. Turnbull makes some good points from time to time which are promptly ignored or twisted as they cannot be argued against without relying on government data ones gut knows to be incorrect. An easier attack on Turnbull if that is what you wish to do… would be to take his own points against FTTH NBN and ask why an NBN is needed at all. However, to make that argument would be just as critical on both proposals… Given that one proposal is clearly preferred by some (I typed idiots but removed it) we will not see such an approach. I look forward to all the captain hindsight’s in 2020.

  18. If Turnbull hadn’t specified “guaranteed 1Gbps” sure he would have been lying, but to have guaranteed gigabit, or something the Gov’t NBN literature would have you believe is uncontended or “guaranteed minimum speed” as Conroy would say, the CVC and AVC needs to be sufficient. The mums and dad’s of Australia think 1Gbps means 1Gbps – it’s the Gov’t’s marketing that needs attention, not Turnbull’s properly defined argument. It’s a debate – and one of the most fundamental factors of a debate are DEFINITIONS. He defined his argument correctly. If that was too intelligent an argument for the average Lateline viewer to follow then we really are doomed!

    Anyway. NBNCo forecast something like 5% of users opting for gigabit speeds by 2028 (from memory, not concrete numbers). Why is the gigabit market even a topic of discussion in the residential user market segment? As part of such a small market these users should be paying their own way, rather than a network being spec’d nationally to deliver the speeds only required by 5% of users.
    In an FTTN roll out, anybody who was willing to cover the cost of a 1Gbps under FTTP would surely be happy to pay the <$5k to have fibre deployed to their home/business.

    To be quite honest I don't know why NBNCo is even putting ads on TV. They're a wholesaler.. it's as relevant as having Telstra Wholesale advertise on TV to mums and dads.. Even less relevant as they're currently destined to be the monopoly wholesaler.

    • If 1gbps speeds are such a distant requirement that won’t be utilised until 2028 and beyond, then why are companies like Google bothering to roll out 1gbps networks in the US? Sure it’s Google who are 1.) crazy enough to try anything and b.) loaded enough to take a hit and heavily subsidise the price. But regardless, they obviously see a need to move as many people onto the fastest fibre connection as soon as possible. Plus the deal comes with a complete cable TV package delivered over fibre, which is undoubtedly the future. One can only imagine that Foxtel and hopefully other content providers pair up with Aussie ISPs in 10 years or so, to offer HD Foxtel over IP + a superfast fibre connection.

      Verizon are also now offering 500mbps a second to FioS customers in the US (admittedly at a high cost) so I think as bandwidth hungry applications emerge for these customers, the benefits of having such a high speed connection will be evident much sooner than most people predict.

      • Google fibre is an experiment to determine just how many people will opt for various services and what use they will put them to. It will give both Google and the greater ICT community tremendously beneficial real world data for planning, projections and investment. Google embarked on that because there is nothing comparable in the US that will give them the data they need. It doesn’t even matter if they take a hit in costs.

        But it is also an example of what is possible – if it is profitable to operate, that will be a powerful argument for other infrastructure providers all over the country. More people consuming more content and services means more money for Google, on the whole – it will lead to new possibilities for innovation and new products and services.

        Google focus is on results and opportunities well beyond operating as a profitable telco – thinking of it in these terms completely fails to appreciate why it is doing it or how it is commercially justifying the project.

        • Fair points. I knew Google Fibre was as an experiment of sorts, but didn’t really know the details behind it. Thanks for the clarification.

      • I recently moved from an Opticomm (non-NBN FTTH) estate to an NBN greenfield estate. In the opticomm estate all free to air TV channels and Foxtel (still need the foxtel subscription and box) streams come in over the FTTH at full signal quality. I move to the NBN greenfield and I have to buy and install a stupid TV antenna on my roof and a powered booster/splitter to get adequate reception. What the NBN ‘should have in 10 years’ is available today… they’re just not interested.

    • I’m sorry but I have to point this out:

      Unfortunately the word lieing doesn’t just mean “a false statement made with deliberate intent to deceive;”, it also means “something intended or serving to convey a false impression”. See the dictionary.com definition

      Turnbull’s statement presented a false impression, even if it wasn’t, by virtue of his wording, a false statement. Therefore it was still a lie.

      • NBNCo can only quote wholesale charges which are completely irrelevant to the end user as their RSP determines the price of their end user service. So the only way to dispute Turnbull’s numbers is to offer an example price of a guaranteed 1Gbps fibre service in Australia (my definition of guaranteed is that I can get 1Gbps speeds whenever I want, every time i want, otherwise the service is operating outside the specifications of the guarantee thus I’m entitled to compensation for not receiving the service I’m paying for).
        As Simon Hackett pointed out in an earlier post, it’s simply not going to be cheap! Whether the $20,000 figure is accurate? Well anybody is perfectly entitled to dispute it with a factual example in practice and for sale in Australia. Turnbull’s numbers are generally based on actual markets, not fanciful figures in Gov’t forecasts.. we’ve had six years of surplus forecasts by the ALP to back up their credibility.

        • Hi there BieRHeDD,

          it is a fact that a 1Gbps residential service from NBN Co will not cost $20,000 per month. If you persist in claiming that it will, I will simply delete your comments. Apologies, but I will not permit irrational debate on Delimiter or falsehoods to be posted. I suggest you read the article again and check your premises.

          Renai

          • >> “both businesses and consumers, could conservatively expect to pay in the realm of $1,200 to $2,500 for a 1Gbps broadband service, based on contention ratios of between 10:1 and 20:1.”

            That’s fine – I would never argue against that, I don’t contest it, and don’t believe I have (apologies if my comments were poorly worded and/or taken out of context). $20,000 is a ludicrous amount of money to expect any ISP to charge for a residential grade service.

            My point is that 10:1 or 20:1 contention on a service should not be confused with a “guaranteed” 1Gbps service. Without highlighting factors like this as Turnbull has done in this debate, voters read election material believing 1Gbps to mean 1Gbps usable in their homes. You only need to look at whirlpool forums for 5 minutes to understand that people without tech backgrounds take broadband speed in marketing material literally. Conroy went to great lengths to repeatedly explain to people that FTTP is a ‘guaranteed’ service whereas FTTN is a ‘best case’ service (dependant on factors such as distance and copper pair quality), regards speed.
            It’s a qualification that should be more apparent but it gets lost in TV interviews such as this one on Lateline because the general public aren’t interested in the tech specs of networks, only what they can get in their house. So Emma Alberici moves the discussion along when it gets too technical.

            I admin a 1Gbps symmetric FTTH connection (residential) in Hong Kong via a major ISP over there, PCCW. I’ve never managed to push it beyond the 300-400Mbps bracket. That’s what the speeds are like on a contended network. There’s no such thing as a ‘guaranteed gigabit’ service if there’s contention involved. I’ve also lived in an FTTH estate in Victoria with insufficient backhaul from estate to POP, diminishing the maximum speeds achievable on 100/40 plans to around 40Mbps during peak times and 60Mbps off peak. This issue was fixed after several months, but the loss of speed did not qualify for any compensation. I was paying for 100/40 on FTTH but i was not ‘guaranteed’ 100/40 even though the final minal was perfectly capable of it. The only way to get guaranteed 1Gbps is to in fact purchase dedicated bandwidth at the kind of prices Turnbull quoted.

            In the interests of the home user (for whom this debate was intended) I think it’s worthwhile for MT to raise points such as this, otherwise user expectations will be that they’ll get 1Gbps from their ‘guaranteed speed’ FTTP, which is in fact contended. That’s where that point made by Turnbull would have led had it not been cut short and moved along to the next question. I don’t claim that Turnbull is infallible, or uses razor thing lines of argument at times, but that’s just a debating technique. This is one topic of discussion where I would have liked it to continue and not be cut short, for the benefit of the average home user to learn about dedicated bandwidth and contention and how you get what you pay for.

            That covers my thoughts on the subject. I won’t persist any further as I don’t have anything further to add, I just wanted to better explain my previous post which was questioned, and qualify that I don’t expect NBN gigabit plans to cost $20k, but I don’t expect them to be guaranteed bandwidth either.

        • As Simon Hackett pointed out in an earlier post, it’s simply not going to be cheap! Whether the $20,000 figure is accurate? Well anybody is perfectly entitled to dispute it with a factual example in practice and for sale in Australia.

          That entirely depends on your definition of cheap. Under the right circumstances I would expect the cost to consumers to be around $250/m, possibly lower if the provider doesn’t mind providing pretty high contention.

          I have already provided solid analysis of how this figure can be practically achieved, and multiple ways to do it. If you feel that you’d rather take a flippant remark from Turnbull intentionally designed to mislead over this analysis on a hypothetical product I have to say you are being completely irrational and acting on profound ignorance.

          • It’s more dependent on my definition of ‘guaranteed 1Gbps’ meaning dedicated 1Gbps. Some may find that an unreasonable definition and to them I apologise. As soon as contention is brought into the equation then yes, $20k is completely false for a single end user, and your $250pm is certainly achievable and apologies if you thought I was disputing that. Turnbull was cut off as he was attempting to go into further detail.

            The Gov’t (Conroy) had been selling FTTH as ‘guaranteed speeds’, which at the wholesale level is certainly true, the network is great and has admirable specs. But I believe Turnbull was trying to back Albo into a corner with a clever (sneaky?) choice of words, to get him on record saying that end users will not get ‘guaranteed’ gigabit speeds because the services will be contended at the retail level to be affordable to households. There’s no other reason for him to add the word ‘guarantee’ unless he was trying to push Albo into a more specific argument about contention. This would have been a great soundbite for the LNP because Albo would be saying that speeds aren’t guaranteed, whereas Conroy had always implied the opposite. The argument was cut off and that disappointed me. Flippant, sure, lie – well in the transcript I can certainly see it that way too but I don’t believe that was his intention, to deceive. If the public had seen that point continue they may be a little more educated about how networks operate and that would have just been super. IMHO.

          • @BieRHeDD

            Whatever Turnbull was trying to do, it was misleading one way or another. It was an outright lie in the way he said it. After analysis you MIGHT downgrade it to misleading, but not for the layman. They see “$20K for ORDINARY HOUSEHOLDS to get 1Gbps”. No matter which way you spin it, that’s a lie.

            On contention- 80-90% of people wouldn’t WANT to know about it and only 50-60% would even understand it. It is a network term and it is a fact of networking. The word “guaranteed” in networking has never been used lightly and Turnbull just threw it in. His intention was to deceive people into thinking everyone would be paying tens of thousands for higher speeds, which is simply plain false. Networking is based on the idea of traffic priority. Turnbull just ignoring that entirely and then getting shirty when Alberici drew him up on how technical it is for her audience (which it IS) is his problem, not the problem of the argument.

          • @ seven_tech
            The only reason I give Turnbull any leeway is because he used the term ‘guaranteed’.. a word which Conroy had used repeatedly to describe the speeds of fibre as a point of difference from copper. I don’t think it’s unreasonable to expect Conroy’s claim of FTTP providing ‘guaranteed’ speeds for the consumer to be clarified in debates such as this. Turnbull has requested another debate but Albo has declined, which is unfortunate because I really want to see their definition of ‘guaranteed’ speeds for the consumer clarified. My opinion is that the usable bandwidth for the consumer is determined by the RSP and their contention ratio and service pricing, not something the Gov’t or NBNCo can dictate.
            I’ve had personal experience with three FTTP connections now, and have experienced first hand the effects of bandwidth contention in fibre networks. The term ‘guaranteed’ being used lightly by Conroy is my gripe. Turnbull was using it correctly, he just f’d it up because he was trying to be slippery, so it came out as a blatant lie rather than starting the straw man argument I believe he intended.
            That’s all i’m giving him… I’m in no way trying to oppose the consensus that in hindsight it was a lie, just analyse why it played out that way.
            Perhaps a better way of doing it would be to simply ask Albo to explain to what extent the speed of data to the consumer will be guaranteed, and if so, by whom. But I’m sure Malcolm got into politics for some debating adrenaline, he just enjoys the grandstanding too much to ask simple questions.

          • I don’t think it’s unreasonable to expect Conroy’s claim of FTTP providing ‘guaranteed’ speeds for the consumer to be clarified in debates such as this

            I think it was obivous, from the context of comparing DSL variants and FTTP, that ‘guaranteed’ in this context means ‘the line speed will be provided as advertised and not affected by external factors such as noise or line length’.

            Yes, maybe Conroy could have clarifed that, but that is a very different definition from “completely uncontented” as Turnbull was making. It was, as you said, a Strawman. And althrough getting it played out may have been interesting, I do not think we should be accepting of delibrate logical fallacies in our political debates.

            And I think that is where we disagree.

          • @NightKhaos & BieRHeDD

            I agree with NightKhaos. Conroy always meant guaranteed as in the line is capable of that bandwidth no matter the length, unlike copper. That is the overall issue and the reason this whole debate exists- the copper cannot handle higher speeds due to the physical and material limitations of the material itself, not the technology powering it.

            I don’t think Turnbull wanted to clarify anything beyond being slippery using “guaranteed”. It slapped him in the face. I think he deserved that. If he wants to discuss contention, discuss contention. Don’t try and base an argument on it and then try and specify you meant by contention.

          • We have that luxury of knowing what Conroy meant because we understand the tech. I agree with you on what he meant – and at a wholesale level it’s 100% true. But to non-tech people (I have brought this topic up with non-tech people for the last couple of years) Conroy saying ‘guaranteed’ speed leads them to conclusions that may not be correct. Speak to a few non-tech people and ask them what they think it means and how it will enhance their broadband experience. Labor’s ads about the NBN being 40x faster than the coalition’s proposal. It’s only 40x faster for 5% of people who are at that speed tier in 2028… This is my issue. I want to topic to be covered in a debate so people are informed. That is all.

          • It’s only 40x faster for 5% of people who are at that speed tier in 2028

            So how much faster do you think they should say?

          • ‘up to’ 40x faster would be fine with me.
            or an asterisk with an example of the comparison that makes the 40x factor.

          • Can you point me to one of these ads please BieRHeDD, Google seems to have let me down and shows nothing about it, but I’d like to see the specific wording for myself.

          • @tinman

            Comment with url must be awaiting moderation… if you go to youtube and search for “Three big reasons to back Labor’s NBN” you should find it… it was posted by ‘ausralianlabor’ youtube account.

          • @tinman

            upon going to the ALP’s iwantmynbn dot com dot au website they use different wording than the youtube video.. there’s our favourite word again.. ‘guarantee’ but this time it’s a liberal guarantee… I wonder if Turnbull actually has ‘guaranteed’ 25Mbps… and then defined guarantee as dedicated (with the $20k comment)… that could definitely come back to bite him unless everybody gets 1:1 from their RSP. :P

            “Here are three BIG reasons to support Labor’s NBN when you cast your vote this September.

            The Liberals #Fraudband can only guarantee speeds that are 40 times slower than Labor’s NBN. And if you want super-fast fibre speeds, you’ll have to pay up to $5000 just to get hooked up!

            It’s now or never for a super fast National Broadband Network.”

          • Cheers mate, I agree, “up to” should be used a lot more than it is for both plans really. Neither party can actually guarantee rock solid “X Mbps” speeds when they don’t control the RSP’s backbone.

            And maybe they’ve been reading Delimiter, they’ve avoided most of the usual complaints people have on their fact sheet (http://www.iwantmynbn.com.au/the_facts), though I’m not sure they are correct when they say “Optic Fibre – Broadband of the 21st century” while coppers is “Last centuries technology”…Optical fibre was invented in the 1840’s and used for computers from the 1960’s :/

          • It’s more dependent on my definition of ‘guaranteed 1Gbps’ meaning dedicated 1Gbps. Some may find that an unreasonable definition and to them I apologise. As soon as contention is brought into the equation then yes, $20k is completely false for a single end user, and your $250pm is certainly achievable and apologies if you thought I was disputing that. Turnbull was cut off as he was attempting to go into further detail.

            That’s the problem, right there. It is an unreasonable definition. There are very few connections in the world, in fact so few we could probably name them all, where a certain performance level is assured to occur 100% of the time.

            Even business grade connections where businesses actually require insane bandwidths are limited and protected by SLAs and other terms, because no one can actually be 100% sure that a connection will perform as stated.

            Then there’s the fact that Turnbull was referring specifically to households. I don’t know about you, but I don’t download 300 TiB a month. That is what a connection at that cost would be equivalent too.

            The Gov’t (Conroy) had been selling FTTH as ‘guaranteed speeds’, which at the wholesale level is certainly true, the network is great and has admirable specs. But I believe Turnbull was trying to back Albo into a corner with a clever (sneaky?) choice of words, to get him on record saying that end users will not get ‘guaranteed’ gigabit speeds because the services will be contended at the retail level to be affordable to households.

            A pointless exercise considering the very same problem applies to his network, and ultimately one used to further muddle the debate. It’s also a bit of a stretch to assume this of Turnbull. You’ve giving him way to much leeway here.

            He presented a false premise, even if that premise was ultimately supposed to lead to a “point” that we can’t be “assured 100% of the time” that we’ll get the speeds promised, it was still a dishonest practice. Something Turnbull does quite frequently, to varying degrees. Calling out Rudd on the “free” claim is one that comes to mind.

            There’s no other reason for him to add the word ‘guarantee’ unless he was trying to push Albo into a more specific argument about contention. This would have been a great soundbite for the LNP because Albo would be saying that speeds aren’t guaranteed, whereas Conroy had always implied the opposite. The argument was cut off and that disappointed me. Flippant, sure, lie – well in the transcript I can certainly see it that way too but I don’t believe that was his intention, to deceive. If the public had seen that point continue they may be a little more educated about how networks operate and that would have just been super. IMHO.

            Actually I can think of a far more simpler reason he’d add that word: to cover his arse in case someone called him up on it. The fact is most people don’t actually know the definition of a lie. They think it means a false statement, when it fact is also means “something intended or serving to convey a false impression”, as I have pointed out multiple times on this article alone. By putting that word in he can rely on this, and you are one of the people who are falling for it, even when Turnbull hasn’t tried to explicitly use the “out” he presented himself.

          • Bringing SLA’s into the discussion is muddying the waters even more. The guarantee was for ‘speed’ not uptime. In my experience with 1Gbps in HK it has never reached 400Mbps or above. That’s why I question ‘guaranteed’ speed on contended residential networks.
            Anyway I’m not that invested in this argument I was just putting forward my opinion of how it played out and that he was never allowed to finish.
            It seems you’re into definitions so check out the definition of ‘debate’. It’s more than a presentation of facts, there’s flexibility and yes the ability to draw on hypotheticals. It’s a strategic method of placing the opponent in a position where it’s difficult to for them to justify their side of the argument. We were watching a debate. It sounds like you don’t like debates. I do. They’re great for exploring topics from multiple dimensions. It’s a performance, an art, an can be very educational. It just happens that Turnbull is very good at them until he gets cut off.

          • Bringing SLA’s into the discussion is muddying the waters even more. The guarantee was for ‘speed’ not uptime.

            Oh, so it’s okay for the company to not provide the advertised speed without penalty? SLA agreements specify performance between sites as well as uptime.

            In my experience with 1Gbps in HK it has never reached 400Mbps or above. That’s why I question ‘guaranteed’ speed on contended residential networks.

            Oh so in the residential space the RSP is responsible for every server on the Internet? The fact is unless it’s a dedicated link between two specific sites there are two many factors outside the RSPs control, which is why they specify SLAs.

            And as I said, Turnbull was specifying referring to households, where this problem of uncontrollable factors increases to the exponential because there is absolutely no way your RSP can assure you of a certain performance unless they know where you’re trying to connect to.

            Anyway I’m not that invested in this argument I was just putting forward my opinion of how it played out and that he was never allowed to finish.

            It was still a dishonest debating tactic. Good debaters don’t have to fall back on convoluted arguments.

            It seems you’re into definitions so check out the definition of ‘debate’. It’s more than a presentation of facts, there’s flexibility and yes the ability to draw on hypotheticals. It’s a strategic method of placing the opponent in a position where it’s difficult to for them to justify their side of the argument. We were watching a debate. It sounds like you don’t like debates. I do. They’re great for exploring topics from multiple dimensions. It’s a performance, an art, an can be very educational. It just happens that Turnbull is very good at them until he gets cut off.

            Oh I like debates, I’m having one with you. Also as for Turnbull being good at them? I disagree. He’s a good politician, he’s not a good debater.

          • @NightKhaos

            It was a straw man argument that played out very badly and yes can be framed as a lie.

            I’m not gonna roll with any discussion of SLA’s because they factor in more properties than speed. I was only interested in discussing and debating the definition of ‘guaranteed’ speed. We can (but I won’t any longer) debate our own definitions here, and the industry definition. Based on Turnbull’s line of argument we can assume his definition is guaranteed=dedicated.

            The only thing I’m interested in is Albo’s definition. Until I hear Albo’s definition, or Conroy’s definition, I have no argument to make against them. If I speak to anybody about the differences of fibre and copper and they mention guaranteed speed i’ll ask them to explain what they understand ‘guaranteed’ to mean in this context.

            No debate to be had here :)

          • >> Calling out Rudd on the “free” claim is one that comes to mind.

            In this greenfields we had to pay the builder $1600 to run a comms conduit from pit to house to NBN spec and for a metal wall cabinet to NBN spec with dedicated GPO inside. Without that NBNCo won’t install. For this reason I question the use of the word ‘free’ too. Off topic for this article though. Just sayin’.

          • You’re right that is off topic, because Turnbull was deliberately confusing up front costs with ongoing. The fact that your property developer didn’t properly conduit your site and decided to pass that cost onto you isn’t NBNCo’s fault.

            The fact is that bad conduit probably would have bit you on Turnbull’s plan as well. If you were promised broadband access in your site you might even have grounds to take legal action against the property developer for undisclosed costs.

    • Guarantee != dedicated.

      I can quite happily offer you a ‘guarantee’ service without providing a dedicated service.

  19. In fact, NBNco FTTP ALREADY delivers 2.488 Gb/s down and 1.244 Gbp/s up to every house connected.
    It is only software in the box on the wall that restricts customer speeds to that of the plan they are paying for.

  20. It’s a shame you attempt to support such a topical argument.

    It is probably true that the $20,000 remark was meant to be deceptive (it would be hard to claim innocence)
    It is false to say that it is a lie (especially given all of your supporting evidence)

    Taking the “any household” remark is as deceptive as the $20,000 figure because it was taken out of context: “which is why, by the way, nobody will buy it other than businesses that need a very big..”

    I think he was pretty clever in being deceptive and avoiding a straight out lie. Unfortunately by the title, we prefer to paint Turnbull a liar rather than his sneakiness and informing people of the facts (as he was making a rather factieos argument to begin with).

    Bad Malcolm for using inciteful figures, bad reporters by turning it into such a piece of drama! We need facts, not grubby politics.

    • As I said to someone else above, and I will repeat it to your here Ralf.

      I’m sorry but I have to point this out:

      Unfortunately the word lieing doesn’t just mean “a false statement made with deliberate intent to deceive;”, it also means “something intended or serving to convey a false impression”. See the dictionary.com definition

      Turnbull’s statement presented a false impression, even if it wasn’t, by virtue of his wording, a false statement. Therefore it was still a lie.

    • Alberici asked, to paraphrase her, “[Would it cost] the average household [$20,000 for a 1Gbps connection]?”
      Turnbull replied “[Yes,] ANY household.”

      The average household wouldn’t use a 1:1 1Gbps connection, but that doesn’t mean they won’t use a 1Gbps connection – period; so you and I both know that Turnbull’s statement was (deliberately) false.

  21. It would seem logical to me that a plan that is 10 times as fast as the current fastest plan would cost at most 10 times what the current fastest plan costs.

    I doubt that there are many RSPs with less than 1Gbps back-haul so no additional bandwidth need be purchased until such time as contention becomes a significant problem. Those with less than 1Gbps probably shouldn’t offer 1Gbps services :-).

    Given more than 90% of customers will be on lower rates, probably about half on the lowest rate, I don’t see a problem.

  22. Let me see if I’ve got this right:
    – Once an RSP has maybe 1000 NBN customers coming via a particular POI, then they’ll be at the point where Gigabit CVC is feasible.
    – Gigabit CVC will enable the RSP to offer Gigabit plans in the area connected via that POI, and those Gigabit plans should cost approx $500 for 1TB quotas.
    – Simon Hackett is looking out for the small RSPs, even though he’s working for one of the biggest.

    Great! Sounds like it won’t take long for an RSP the size of the iiNet group to be able to offer 1000/400 plans to the majority of customers.

    • Exactly, and if the small ISP wants to offer it, they just have to try some “out of the box” thinking, like Google Fibre inspired “minimum commitments” in PoIs.

  23. This has probably already been mentioned (haven’t read the comments yet, cooking dinner!) but it’s not JUST contention ratios that will make the costs much MUCH cheaper than $20k per month.

    Each retail plan will undoubtedly be limited to something like 500Gb-1Tb of downloads per month. Which means you could probably have contention ratios much higher than 10-20:1 eg. 100 or even 200 retail users sharing that gigabit pipe. That’s crazy! I hear you say …well not really. Unless these 1gbps customers are on unlimited plans they WILL NOT BE DOWNLOADING CONSTANTLY or they would quickly exceed their quotas.

    With a 1Gbps pipe you can transfer 300 TB of bandwidth in a month (roughly, iirc). Now if the average retail customer’s monthly limit is 500Gb, I don’t think there’d ever be that much issue with 100 home Gbps connections sharing a Gbps connection.

    The higher the speed allowed the higher the contention ratios can go, especially as the monthly download limits are not going to scale up proportionally to downlink speed (eg. 100Mbps vs 1Gbps connections)

  24. Or another way to put it … ISPs could easily sell 1Gbps cheaply if they simply lower their monthly download limits to eg. 100gb or 200gb a month for these types of connections. This would still suit the vast majority of people wanting a 1Gbps connection because most of them aren’t in it to download terabytes of crap, most just want a superfast connection.

  25. $20k a month is low estimate. You can’t save on contention if you don’t have anybody to share the line with. On the NBN you can’t share across the entire network, only across a point of access. This means only a handful of points of access will ever get the investment for 1Gb, where it is commercially worthwhile.

    When the first person at a POP signs up for 1Gbps, lets say somebody in Kiama, their access provider will have to go out and purchase a full 1Gbps of backhaul, connect it up and then get 1Gbps of bandwidth, which is likely $50k or so retail. Even with 10 people signed up at a POP for the 1Gbps service it is still $5k per month (likely much higher – for eg. Wollongong couldn’t handle 1Gbps on its backhaul). Then when you add in other providers, it exacerbates the problem.

    If you email the service providers and ask them where/when they will offer 1Gbps, you will see this problem. There is an investment problem because most customers are going to be happy with 50Mbit or 100Mbit so you don’t have enough customers to share the 1Gbps costs across. There is a reason why every FTTH network in the world is either 100Mbit on existing backhaul or 1000Mbit on new backhaul, and not 100Mbit upgradable to 1000Mbit – because of this problem (you essentially kill the commercial viability density)

    The 1Gbps plans were a hack on top of an NBN plan that was well considered to work with existing backhaul infrastructure. It was a political response to a technical problem, because the debate suddenly became 100Mbit vs 100Mbit and the NBN had to do something to prove their network was better or more capable. It is a plan used for marketing, it isn’t feasible for most homes.

    • hey Nik,

      I approved your comment because of its general usefulness to the debate. However, I will note that in this statement: “$20k a month is low estimate.”, you are referring to the cost to ISPs — not the cost to consumers, which is what Turnbull was asked about. The retail cost to consumers of a 1Gbps service on the NBN will never be $20k per month. That is not a cost the market would ever bear, and it’s not what retail ISPs will charge. Please be careful how you phrase this in future.

      Renai

      • Except Renai, you missed the crucial word that Turnbull used – guaranteed. That is, in service provider world to offer a customer 1:1 contention ratio. So the statement Turnbull made is 100% correct.

        • Adam and Mark:

          I’m sorry but I have to point this out:

          Unfortunately the word lieing doesn’t just mean “a false statement made with deliberate intent to deceive;”, it also means “something intended or serving to convey a false impression”. See the dictionary.com definition.

          Turnbull’s statement presented a false impression, even if it wasn’t, by virtue of his wording, a false statement. Therefore it was still a lie.

          • Actually no. The only way to achieve guaranteed throughput is 1:1 contention ratio. If you go and look at the premium providers and the terms of guaranteed throughput, the only way to do so is with no contention. Dedicated isn’t the word used. Guaranteed is. Might want to listen to those with actual enterprise industry experience at a service provider.

          • And that “guarantee” is usually made via a SLA which means that your service is probably moderately contentend, depending on the SLA that could be as low as 3:1.

            No, NPSF3000 is correct, unless you’re using dedicated dark fibre, there is no way to be absolutely assured of 1:1 contention.

          • @Adam

            I think you’re getting into semantics, but if you really want to:

            Dedicated: Exclusively allocated to or intended for a particular service or purpose.

            Guaranteed: Provide a formal assurance or promise, esp. that certain conditions shall be fulfilled relating to a product, service, or transaction.<i?

            Dedicated suggests fibre allocated specifically for you. Guaranteed is much more general and talks about an “assurance” or “promise” to provide a certain service- ie 1Gbps.

            I have seen BOTH used in the context we are talking. But I’d consider the second considerably closer to reality. Very, very, VERY few enterprises would have their own dedicated dark fibre. But they WOULD have a guaranteed bandwidth as part of their SLA.

          • I think you will find companies with guaranteed 1:1 contention do not actually provision 100% of the band width required for everyone to do so simultaneously, because in real life it doesn’t happen. They provision enough to cover the customer base being able to run at full speed. Since not everyone is not going to be doing so at once, they provision what they need and increase it if demand increases.

      • Apologies, I thought referring to the ISP cost was implied since I was speaking about their contention ratio, backhaul costs. etc.

        My main point on that topic is that even with 1 gigabit being activated this year nobody is going to be able to provide it (except possibly Telstra). It was planned to grow into backhaul and rollout capacity, and it will still be ready when the market and backhaul is ready (you can’t have a feasible 1Gbps service when the rollout is at 5, 10 or even 30% – you need 2-3% adoption with a complete rollout, and that doesn’t take into account other service providers as competition).

        Also re: your calculations on a 1 Gbps – is that not just the AVC cost and CVC? Wouldn’t you need to add in NNI (negligible), backhaul (CVC gets you to the next suburb, backhaul is what serves those javascript polls of facebook.com), a retail margin and then GST?

        The smaller providers (outside of Telstra and Optus) are going to have a tough time finding backhaul from some of the more remote POI back to the Internet.

        In buying colo, I know what those prices look like – and more pressing at the moment is the lack of supply, it is really really difficult to purchase a gigabit in large parts of Australia at the moment. This is why the gigabit was planned for later, so that the network could catch up with what the NBN was going to dump on it.

        Further, since leaving that last comment though I had a chance to watch the interview. I think the reason why Turnbull is being misinterpreted is because he was cut off. One of the last words he said before he was cutoff was ‘contention’:

        MALCOLM TURNBULL: No, no – that is certainly not true. If you – when you are a retail service provider, you’re Telstra, Optus or whatever, you have to buy a pipe of certain capacity from the customer’s – your customer’s premises back to the exchange, the point of interconnection. And the size of that pipe will determine how much contention your customers have.

        EMMA ALBERICI: I think we’re getting a little more technical – too technical for the purposes of this discussion because …

        MALCOLM TURNBULL: Well, you always shut me down when I’m about to make the most important point.

        I think he should be asked what the remainder of his point was, and given the benefit of the doubt – it seems like he was about to explain what he said in the first part as being an ISP cost component (its not like ISP cost and user cost are totally unrelated).

    • @Nik

      The 1Gbps plans were a hack on top of an NBN plan that was well considered to work with existing backhaul infrastructure. It was a political response to a technical problem

      This statement is simply false. NBNCo’s original Corporate Plan in 2010 projected 1Gbps plans being offered by 2015. It may have been political to bring it forward. But it is technically achievable, useable and useful for those businesses and bleeding edge consumers in the fibre footprint.

      NBNCo. plans to upgrade to 10GPON by towards end of rollout, running concurrently with GPON, giving a combined throughput of 12.5Gbps down and 3.75Gbps up between 32 users. Your assertions about backhaul are unfounded- NBNCo. have a design limit of 70% capacity before adding more backhaul.

      I can gather then you will be surprised to see, if the NBN is allowed to continue, that a, granted small, but significant % of businesses will likely take up 1Gbps plans that RSPs WILL offer early next year. Compared to tens of thousands a month for a leased fibre line, a few thousand for a contended pipe is a godsend.

      • I know, and IIRC they even published the $150 AVC for 1000/400 in the original 2010 doc. I’m talking about bringing it forward, the backhaul network and economics of providing it were only going to work when they were going to work, the NBN moving the date forward isn’t going to change that. 2015 was/is too early, so there was little point in upgrading the ports now and promoting it as being activated in 2013 when nobody is going to pick it up (except perhaps Telstra).

  26. Really a moot point from Mal saying 1Gbps will cost $20K…

    Because, no one will ever need more than 50Mbps via those iron (oops I mean copper) wires anyway ;)

  27. Just thinking about this the last day or so, but what sort of contention ratio will bring the cost down far enough for the home user to be happy to pay?

    With all the talk of $20,000 or $430, or whatever dollar amount is tossed out there, I cant see anyone being happy to pay more than $100 a month for their internet, give or take a few bucks. Will there be exceptions? Sure, there always is, but for 95% of the population they wont want to pay more than they currently do.

    So in 5 years time, when our needs outgrow the 100 Mbps plan, is that a Gbps pipe still going to be costing $20,000 for 1:1 contention? I sure hope not, because that is a big kick in the guts for everyone.

    • @GongGav

      A 1:30 contention on ADSL and consumer fibre now is fairly standard. 1:30 for 1Gbps would cost $150 AVC + $660 CVC. I think there would be a higher contention on 1Gbps plans, as 75% of the pipe couldn’t be used for most servers now and probably 50% in 5 years time. so a 1:75 is probably closer. That’s $150 AVC + $260 CVC. You might even push it out to 1:100 to keep the cost under $500.

      Until we start exceeding NBNCo’s thresholds for bringing down wholesale prices, 1Gbps is going to be a niche business and bleeding edge consumer market. 250 and 500Mbps are more likely for high speed consumer plans due to the reduced AVC. I could easily see a 250Mbps plan being sold for $200/month with a decent few TB quota.

  28. An interesting and perhaps overlooked concept is that a consumer fortunate enough to already have an FTTP connection will also have an NTU on their wall in the premises that has 4 data ports, and each can be provisioned with an RSP subscription from a different ISP at different speeds and monthly data allowances.

    For a share-house with multiple unrelated residents this allows up to 4 of them to subscribe to their own connection, a rather beneficial facility if required. And more profitable per premise, for NBNCo and the retail market.

    Also the FTTP NTU may instead facilitate an even better option for an SMB or home office context in that 2 ISP subscriptions can be made from a couple of the ports on an NTU, one subscription at 1Gbps/400Mbps and another at a slower speed of say 50Mbps/12Mbps, but in this case with an unlimited monthly data allowance, then with careful planning and configuration, there would exist the facility to have unlimited data available at lower speeds and also the capacity for ultrafast uploads and downloads over the alternate 1Gbps service when such speeds were specifically needed, with adequate usage metering on the limited data allowance service a user could really have their digital cake and eat it too. Simples!

    Unfortunately, the Fraudband NBN will not provide these benefits, as the most you may get now is a single FTTN port and modem that supports only one account that also reduces the multiple subscriber profit that NBNCo may otherwise have made from some rental premises.

    Malcolm and Tony don’t think we need anything faster than ADSL2+ speeds duplicated on FTTN and we don’t need to ever have any fast upload capacity, for if we did, intelligent ideas will spread throughout the community much too fast, and the luddites will be undone;-(

    As to the concept of costing $20K CVC & AVC per month for the synchronous NBNCo 1Gbps service, Goggle Fibre currently charges US$70 per month in the USA for unlimited data, or US$130 per month with unlimited data and multi channel TV with many other beneficial services. Parts of Austin, Texas are also going to 1Gbps synchronous on AT&T next year, cost undecided, but expected to be competitive with Google Fibre.

    In France today, an ISP announced that they are increasing the speed of their FTTP uploads and downloads, at no extra cost to their customers, this service is also moving to 1Gbps synchronous within a year, probably at no extra cost.

    Therefore the whole economics of the NBNCo wholesale charging model must be well out of step with the overseas models for billing, and the form of the backhaul construct and contention provisioning must really be akin to the type of early IT assumptions made that led us to the Y2K debacle.

    Something fishy somewhere, about the whole concept of the NBN if we can’t match the costs and benefits of France and the USA’s versions of FTTP.

    Australians really do get taken for a bunch of mugs on service and product costs. And none of the politicians care, because when we pay such high prices, the tax man always gets his 10% GST cut. Therefore it is always in the Government of the days’ interest to keep us paying ever more for everything.

    No wonder so many families are actually struggling behind the scenes. But my concept above for maximising the benefits of 2 types of ISP services on the one FTTP connection would be the type of smart strategy that at least gives small business people the scope to wisely invest in their own growth potential.
    Let us hope that the COALition version of NBNCo retains the same 4 port NTU for FTTP on demand installations.

    Otherwise the South of France is starting to look real good.

    Illegitimi non carborundum.

    • I suspect most typical share houses would find the single shared account approach to be cheaper for the household as a whole. But the class of people who might benefit from the examples you give would number a smaller percentage of the whole. That being the case, shouldn’t a 4 port NTU only be offered as an optional, user pays extra?

Comments are closed.