The Pencil Guy: Hourann's illogical blog

Fibre at what cost?

Reader response time: an e-mail I got yesterday pointed out that at no point in the Prime Minister’s announcement did he talk about how much people would pay to get shiny new FTTH connections! And presumably this will be a significant detail for any private investor deciding whether they want to jump into the venture. In fact, cost questions seem to be the crux of most of the criticism I’ve read (not counting fluff from people who don’t understand that upgrading speeds on fibre is relatively easy; just because it’s built at 100Mbit doesn’t mean it’ll stay there).

Thus I thought I might do a quick and thoroughly unrepresentative survey. In Hong Kong, it’s HK$600 per month for PCCW 100Mbit fibre (that’s US$75 or A$100-ish), while Bredbandsbolaget in Sweden charge 320 kronor (US$40 / A$55) for the same. In New York, Verizon will slug you $140 (A$200) for 50Mbit or $60 (A$95-ish) for 20Mbit — though note that their base US$45 offering is worse than my iiNet connection in Perth!

However, none of those networks was built with government money. South Korean fibre got about $1.5 billion worth of subsidies in the early 2000s (mostly for rural areas), and a Korea Telecom 50Mbit connection in Seoul costs a mere 26 000 won (US$20 / A$27) per month.

Of course, prices in Australia would never start that low. But then, my dial-up connection in 1996 was $5 per hour of use (thank you, Telstra), and of course prices will drop over time. Early adopters may well be slugged $100+ a month to cover construction, but given that the network will be owned by a wholesaler that’s not likely to remain a problem long-term. Optic fibre won’t go obsolete anytime soon; a properly-managed FTTH network should hopefully mean cheap and fast Internet five or ten years after it’s built …!

  1. Hey man,

    One thing to keep in mind is that it’s actually FTTP – P being “premises”. So if you live in a duplex/triplex/strata plan/etc. – you get one fibre to the land/lot – not one per dwelling. Which is kinda short-sighted IMHO.

    Also, was going to point you at https://lists.sage-au.org.au/pipermail/sage-au/2009-April/029530.html – but then I checked and found that the SAGE-AU archives are member only :-/

    So without giving too much away, Mark Newton’s analysis crunches some numbers…he’s an engineer for Internode and is generally regarded as “knowing his shit” :-) The most interesting part is the conclusion, based on some fairly conservative estimates:

    [snip...]

    So on a $100 per month service fed over a $70 per month NBN local loop, the ISP would end up assuming that the customer’s 24×7 average will be about 100 kbps, which translates to about 25 Gbytes worth of downloaded data, give or take.

    So on those 2009 numbers your $100 per month will come with a 100 Mbps NBN port and a 25 Gbyte quota. How’s that look for a retail plan?

    There could be significant variation if the $70-odd per month for the NBN access tail ends up being different. It could be different if the cost is shared between an ISP, a phone operator and a cable TV operator. It could also be different if the Govt never expects the NBN revenues to pay for its cost of deployment.

    [snip...]

    I’m currently on a 25GB plan. Regularly get shaped at the end of the month. It doesn’t cost me $100/mo…if I had fibre, I reckon I’d chew through 25GB in a lot less than a month; I have actually done 19GB in significantly less than a day once. The rest of the month was tough :-)

    I think fibre FTTH has to be a part of any long-term plan, and kudos where due to the Govt. for recognising that. However, that doesn’t make the current proposal any less backwards, ill-considered and very much a knee-jerk reaction to their realisation that the NBN v1.0 proposals were all pretty much crap.

    I also think describing any part of Korea as “rural” [especially when comparing it to "rural Australia"] is somewhat disingenuous, but I won’t get into that :-)

    M.

  2. On the FTTP / FTTH distinction: fair call, but none of the places I’ve seen have fibre all the way from an exchange to individual apartments (not that it doesn’t exist, it just seems rare). And the percentage of Australians in detached houses is waaaaay high anyway :-P

    On $100/month service, that analysis sounds perfectly valid. But I do not envisage $70/month will still be the wholesale cost a few years after deployment, in the same way that wholesale prices for 3G phone service are down on where they started off (enabling things like Virgin Broadband, for instance).

    And on whether it’s a knee-jerk measure, so what? Most people who bagged the earlier plan said “FTTN bad, FTTH good”. So now we’re getting FTTH, and the problem is …?

  3. The problem is I probably redacted too much from the original email :-(

    The opex cost of “Newer Shinier NBN” is actually far more significant than people are considering ATM. The number crunching was based on the cost of delivering mebabits from CA to Sydney. And this was done with 2009 prices, with allowance for previous trends, and the $70 figure included allowances for “drops in the wholesale price” and amortisation of business expenses, capex and other opex over the next 5-10 years.

    Nutshell? $70 will be the _lowest_ price per month…unless the Government never expects any return on its investment or expects the private investors to go bust.

    Knee-jerks in this industry are scary when they come with a big fat rope connected to Conroy. Yes, he really is as incompetent as everyone thinks he is. Plus there’s always the distraction factor [i.e. are they doing this _now_ to distract us from something else they don't want us to pay attention to?]

Care to leave a comment?