Friday 16 August 2013

NBN: Unanswered business questions of the Turnbull Node Plan

My current list of business, not technical, questions about the Turnbull Node Plan.
There are some questions about implementation details and things like "Traffic Class" that arise because they're on the NBN Co rate card.

From what Turnbull has published, we know, more than merely infer, that Turnbull did indeed prepare a full Business Plan, out to 2040 at least. A reasonable Business Plan would include the replacement of the Node (FTTN) network as described multiple times in the Coalition documents.

This is at least a $4 billion charge to the full project, referred to by the Coalition "CapEx Reuse", normally this would be classified as "deliberate waste".

This is, by my calculation, the whole of the CapEx saving achieved by the Turnbull Node Plan, the $17 billion claimed is impossible, even if the full Telstra payments are swapped to OpEx in a "rent, not buy" deal - leaving the original contracted "PSAA" payments of around $7.5 billion still to be paid later. This point was raised directly with Turnbull's office very early on and has never been explained.

The Parliamentary Budget Office, PBO, has sufficient capability and time to evaluate a full spreadsheet model of the Turnbull Node Plan (if supplied electronically) and to extract the financial/marketing assumptions and refer them to experts in DCBDE (Department of Communications, Broadband etc), NBN Co or external Telcos like Telstra and Optus, or the many experts & consultants in the field.

It doesn't take long, a day or two, for real domain experts to assess given assumptions, like traffic growth and market demand. Interest rates and A$ exchange rates are well within the expertise of the PBO. A panel assessment of the Turnbull Node Plan would allow PBO to quickly create a credible range of estimates. There are good techniques, such as "Delphi", to apply these estimates.

Outstanding Questions on Coalition Business Plan:

First, a competently assembled Policy with adequate supporting documents would have included the full Business Plan to 2040. If the Coalition has intended to clearly annunciate its Plan, for its communication to be be complete and clear and for its assumptions to be open and transparent, it would have gone about things very differently. Instead, that Master Magician, Turnbull has engaged in misdirection, distraction, obfuscating, deflecting, confusing & conflating, attacking, inverting positives ($110 ARPU is "bad") and drowning all an sundry is masses of superfluous and irrelevant detail.

This is why the PBO cannot examine the Coalition NBN Policy and would need a huge amount of expensive, external consultant time ($250,000-$500,000) to properly assess the Coalition documents: they are not designed to convey pertinent information, but to prevent exactly that.

That the Coalition haven't submitted their full Business Plan to 2040, and that they've deliberately withheld critical and necessary headline figures, like ROI and pay-back period, speaks volumes in the political world. It's as close as it gets to a flat-out admission of deliberate action.

They are presenting a Plan that will fail, they know will fail, is designed to fail and will complete Turnbull's commission to "Destroy the NBN" by sending it bankrupt.

As Turnbull has started to state frequently now, "nobody has challenged my figures in 4 months".
Which raises the question:
Why has the Mainstream Media not asked and had answered all these questions in the last 4 months?
  • What's the CapEx, in their model, of the FTTN component to 2016 & 2019?
    • What is the CapEx, in their model, of the "90% premises to 50Mbps" upgrade between 2016 and 2019?
    • Is this upgrade via Vectoring, increased nodes (shorter distances), a combination or something different?
  • What interest rate is assumed in their model? per year, if it is modelled as changing.
    • what cumulative total, by year, is borrowed in their model?
    • If NBN Co is not cash-flow positive in any year, how will they fund that?
  • What is the assumed project life of the FTTN component?
    • Do they assume a "straight-line" depreciation in their model?
    • What is the ROI, in their model, of the FTTN component?
    • Does that include the "50% Capex _not_ Reused"? [pg 14 of "Policy Background"]
    • What's the break-even period in their model?
  • The new "Statement of Expectations" to NBN Co include non-commercial directives about prioritising rollout to "poor service" areas and "Cost-effectively as possible".
    • In their model, is there an extra cost for the prioritisation?
    • When will the Coalition more precisely define for NBN Co the meaning of "Cost-effective" in this context?
  • What are the OpEx payments to Telstra, in their model, of the FTTN?
    • How are they modelled to change over time?
  • Is the FTTN component a "pure digital" network or Telephony with ADSL2/VDSL2 on top?
    • Does the Coalition FTTN model follow the UK model of ADSL2/VDSL2 only in the nodes and telephony routed back to existing exchanges?
    • Does their model allow for service disruptions for telephony and ADSL to consumer on cutover to Nodes?
    • Under the FTTN model, will consumer ADSL2 services continue to operate normally after cutover to a Node? Allowing for a disconnect/reconnect?
    • What is the maximum service outage planned for in their FTTN modelling?
  • What does the figure of "8,968,000" on pg 7 of the Plan (finish of rollout 2019) refer to?
    • premises "covered", passed, ports built or something else?
  • If the 8.968M figure is premises passed or ports built, why is that number so specific?
    • Does that imply the FTTN rollout will only occur in the major urban areas?
  • The Coalition Plan states that by 2016, both at least 25Mbps 'by late 2016" will be available "to every household and business" and that ~65% (5.829M premises) of the FTTN rollout is expected to be completed by 2016-17. If there are currently 2.5M premises covered by HFC and 3.139M premises not to be covered by FTTN by "late 2016", how will, in their model, the 639,000 premises later to be included in the FTTN footprint be connected with 25Mbps?
  • Does the FTTN component in their model assume compensation payments to owners of DSLAMs with residual-values, ie not fully depreciated?
    • Does the Coalition model treat cutting the copper back to the exchange as different to offering an alternative broadband service, Fibre or Wireless? If so, on what legal advice?
  • Will the Coalition offer a VDSL2 "NTD" on the same terms as Fibre and Fixed Wireless?
    • If so, will there be any NBN Co charge to customers or retailers?
    • If not, who will pay for the provision, installation and testing of a VDSL2 "NTD" for a customer?
  • In their model, what are the initial FTTN AVC charges per month?
    •  In their model, what are the AVC charges for HFC, Vectored VDSL2, ADSL2 and VDSL2 services?
    • Does the Coalition modelling to 2016 include income from the 35% non-FTTN connected premises?
    • In their model, what are the AVC charges for FTTN & HFC in the period 2016-2019?
  • Does the Coalition model for the FTTN component include the 10% decrease, in real terms, "over the next 10 years"?
    • Is that date in 2013, 2014 or 2016? Is it 1-Jan, 30-Jul or "late in the year" (up to 31-Dec)?
    • Does that 10% decrease apply across all FTTN and HFC AVC services,including Vectored VDSL2?
    • Does the one AVC wholesale price reduction apply to all NBN Co services, Fibre, Fixed Wireless and Satellite, or just the FTTN/HFC components?
    • What are the long-term AVC charge reductions modelled by the Coalition, to 2040?
    • How does this "10%" compare to the already stipulated ACCC AVC price reductions and the price reduction curves published by NBN Co on 19th April, 2013?
    • Does the Coalition model follow the existing CVC price reduction
    • curve, starting at 120GB/mth/avg download, already published by NBN Co?
    • If not, what is the reduction plan and trigger volume for FTTN component and whole NBN Co traffic?
    • - What is the average consumer data download volume (GB/mth) included in the Coalition model for both the FTTN component and the full NBN for all years out to 2040?
  • In their model, what proportion of FTTN services have a telephony service?
    • What are the wholesale charges for Telephony to Retailers in their model?
    • Is this solely a line access charge, or are there other charges (time, distance, ...)?
    • Is this charge, like Fibre and Wireless included in a standard service with a rebate for non-use?
    • Will telephony traffic from Nodes use the same TC1 [traffic class 1] switching network as Fibre?
    • Will customers be able to access TC1 telephony services from their in-premises equipment, or forced to use TC4, normal data, VoIP services?
  • Will the Coalition FTTN and HFC networks allow customers access to the full range of NBN Co layer 2 bitstream services?
    • direct IP (IP over Ethernet), not soley PPPoE used by ADSL2.
    • multicast at TC2 or TC3.
    • TC1 services
    • multiple "QinQ" VLAN services at customer premise
  • Does the Coalition modelling to 2019 include loss of NBN Co traffic across all delivery networks due to local competition & price under-cutting by them?
    • Does the Coalition policy demand national players like Telstra and Optus charge a single national price?
    • Require them, if they choose to undercut NBN Co pricing under the "ACCC price is a cap" change, to offer the lower prices everywhere they can offer the same service?
    • How is this policy on "price cap" not the major Risk factor identified in the 2010 NBN Co Corporate Plan, "Cherry Picking"?

Sources:

NBN Co:
http://www.nbnco.com.au/assets/documents/nbn-co-corporate-plan-6-aug-2012.pdf
http://nbnco.com.au/assets/media-releases/2013/report-to-parliamentary-joint-committee.pdf


Coalition Papers:
http://www.malcolmturnbull.com.au/assets/Coalition_NBN_policy_-_Background_Paper.pdf
http://lpa.webcontent.s3.amazonaws.com/NBN/The%20Coalition\U2019s%20Plan%20for%20Fast%20Broadband%20and%20an%20Affordable%20NBN.pdf

2 comments:

  1. Good questions Steve, I have a couple of comments

    1.
    "What is the CapEx, in their model, of the "90% premises to 50Mbps" upgrade between 2016 and 2019?"

    In Malcolm's background paper he shows a NPV comparison between FTTN and FTTP and makes the claim FTTN is better even at the three year mark (ie 2016).

    But what about the additional capex for upgrading the remainder of the 90% fixed line fottprint to 50Mbps? What happens if the NPV comparison is run out to end of 2019 when that upgrade capex should be accounted for?

    2.
    Regarding the fttp, fixed wireless and satellite rollout, will a LNP change any of the business side of that roll out? I understand the build will proceed as currently planned on the technical side, but will a LNP require subscribers to pay for their NTD and standard installs?

    ie Saying that there will be equivalence between FTTN and the rest of the network does not guaranteed subscribers will not be charged for their NTD or its installation.

    3.
    When the 50Mbps to 90% of fixed line upgrades occur between 2016 and 2019 will subscribers need to buy a second new (first one for plain VDSL2, second one for G.fast/vectoring/fairy tech) modem to get faster speeds than originally obtained?

    ReplyDelete
    Replies
    1. Roo - thanks for comments and suggestions.
      All valid. [and please post others as you think of them]

      The Turnbull Node Plan is woefully lacking in detail and omits so many critical details as to be truly astounding...

      1. Look _again_ at that supposed NPV on page 14.

      - it FAILS to mention both sides of the Ledger, Revenues and Expenses. NPV's are used to compare _investments_ not costs. Investments make _profits_.

      - it stops after *4* years, not by 2021, not after 20 years (2034), not by 2040
      Why? even on his OWN biased figures, if he runs just ONE more year, FTTP is cheaper.

      - You're right, every rock you turn over contains a nasty surprise; The multiple upgrades are UNCOSTED.


      2. yes, I've been making exactly this point for a very long time: At the VERY best, Turnbull's FTTN can only save $4 billion, not $17 billion as claimed.

      We've no information on changes to FTTP rollout, like charging for NTD's and standard installs.

      The Node Plan also creates a back-end loaded upgrade where they INTEND to throw away the FTTN, which by that same NPV on page 14 you quote, WASTES 50% of their investment...

      He saves $4 billion and wastes $4 billion, in real prices. THATS NOT A SAVING.

      The rest of it, no NTD's (not Apples and Apples), and CapEX/OpEx costs forcefully transferred to subscribers only increase costs.

      3. The reading I've done on Vectoring suggests 2nd gen equipment can cope with "non vectoring compliant" subscriber modems. Ie. one old modem doesn't kill performance of every other service.

      That still creates a problem. Vectoring, as noise cancellation over a complex matrix of interconnected signals, _requires_ vectoring modems to provide any subscriber improvement.
      - that's another impost on the subscriber: buy another modem
      - those "leading edge" modems will be more expensive. bigger, more powerful chips, newer (= premium priced) and lower-volume (= higher cost recovery)

      G.fast is pure fantasy. 1.25M nodes and 1-2MegaWatt radiated RF pollution PER CITY (at specified 1Watt/device). At very best, it's interesting in very particular circumstances, not general deployment, but 1Gbps ethernet over cat-5a/6 cable would beat it hands down in internal installs.

      Delete