Monday 15 April 2013

NBN: Correcting the Record on the Coalition NBN Plan

This is a collection of Omissions, inaccuracies or misleading statement in the Coalition's NBN Policy documents.

Please Note:
  • The comments here are concise and assume a more technical knowledge than "an ordinary reader".
  • I'm NOT supporting either Political Party or their Policy.
    • Both Plans have deep, concerning issues.
    • Both side of Politics are making unfounded statements and questionable promises.
  • This piece is NOT analysis or recommendation for or against the Coalition Plan. It only seeks to layout facts that can be checked and confirmed on the public record.
[Update 16-Apr-2013: U1]
  • The Three Elephants aren't discussed by the Coalition. [U1]
    • Significant Pent-up Demand won't be satisfied by DSL.
      • Yes, Tony, speed doesn't matter, but only when its adequate.
      • The ABS is finding Australian Data Volume growth still very high (50-60% CAGR).
      • The Hong Kong experience is when arbitrary constraints imposed by network are removed, the market ramps up then plateaus. This is the usual economic model for free markets: they mature and then grow moderately.
      • The Coalition would like to pretend that Fixed-line markets are different from Mobile Broadband driven by smartphones since 2007.
        • Mobile Broadband use, 3G & 4G, has exploded because of one breakthrough technology.
        • It convincingly demonstrates that consumer demand for the right Telecomms services is highly elastic.
          • The utility of high-function Internet services is very high to consumers.
          • The ARPU growth in mobile broadband will translate to fixed-line services when the right services are provided.
      • The DSL technology with limited upstream channel and limited share capacity per node will introduce significant network congestion and resource/link contention well before expected with simple arithmetic.
        • The Gungahlin Experiment (ADSL via RIMs) definitively demonstrates this.
        • The upgrade cost for active equipment in 60,000 nodes and 20,000 ISAM's in Exchanges will be prohibitive. What we get first up is what we'll be stuck with.
    • Video Streaming is the modern Pay TV
      • Whilst mentioned on Pg8-9 of Background docs, modelling of the impact is omitted.
      • The main beneficiaries of Video Streaming are not the service providers, but Content Providers.
      • Google, via Youtube, is set to enter the Paid-Video market.
        • The Coalition ignores this underdeveloped market completely.
      • Nielsen reports Australians watch ~100 hours/month of TV and 47% have PVR's, with the use of time-shifting growing at 25%/year.
      • ThinkTV reports: 63% households have more than one TV but only 29% subscribe to pay TV. Their Insights Report suggests the Free-To-Air extra digital channels are providing more variety and desirable content to Pay TV households, with FTA now the majority content watched at night in Pay TV households. Pay TV subscriptions are also in decline.
        • With affordable, reliable broadband of any type, 100% of broadband households will be potential Pay TV subscribers, both to current high-cost content providers and New Media providers.
        • We know from the iTunes and ePub industries that there is very high consumer demand for zero- and low-cost ($1) digital goods.
        • Aggregators like iTunes, Amazon and Google (Youtube) are the mediators of this new business. The demand and turnover for small-vendor, low-price content is massive and growing.
      • If 40% of households shift their TV download to Broadband over the first 5 years then Data Volumes will be substantial (~4 times current use added):
        • 40% @ 8.5M = 4M households with 5-6M TV sets & 5M PVR's.
        • 3 hours/day @ 3Mbps StdDefn (SD) is 9Mbps-Hrs/day or 108GB/month additional demand for these subscribers.
        • 3 hours/day @ 8Mbps HighDef (HD) is 24 Mbps-Hrs/day or 288GB/month additional demand for these subscribers.
        • An additional demand of 350GB/month for 40% of subscribers is an additional 140Gb/mth increase in Average Data Volume.
          • Compare to current Average Data Volume of 20-30Gb/month.
          • There is a strong correlation between line access rates and Data Volume.
          • Higher speeds lead to more use of the services.
    • Systematic and long-term Coalition under-investment in critical infrastructure, including Telecomms.
      • Frank Blount, CEO of Telstra, is reported  to have written a report in the early 1990's estimating that the PSTN would be shutdown by 2010. That was a reasonable technical view and could've been implemented well within the
      • In 2005 Trujillo went on record as saying that Telstra had under-invested by $2-$3 billion in both CapEx and OpEx. This proposition was probably only meant for the Government, not for public consumption.
        • If he admitted to this amount, the actual shortfall was probably between $5-$10 billion.
        • The situation has only worsened as staff cuts via redundancies and "Milk the Cash Cow" strategies aggressively implemented to sustained revenue from the PSTN.
        • The Telstra shareprice, the valuation by investors, reflected this until 2011.
      • By 2002/3 the PSTN was in decline, yet the Coalition ignored the mounting problems and did nothing.
        • The decline of the copper PSTN and the need for full fibre, inherently digital, had been forecast and examined since the early 1980's.
        • "Digital Convergence" was known and recognised as a trend since the mid-1980's. This was known to be a driver changing all media: Phone, TV, print.
        • When the Internet/Web broke through in 1996, it was obvious it would be the technology that would drive "Digital Convergence" and to achieve this technical endpoint, only Fibre would serve for Fixed-line connections.
        • The prodigious and long-running super-exponential growth in Internet Data Volumes was well known and studied:
          • This was the direct cause of the "Dot Boom" in year 2000.
        • Yet the Coalition government from 1996-2007 not only did nothing but deliberately ignored sound commercial advice on the matter. Trujillo put his own neck on the line 5 weeks after starting in 2005, delivering a plan to change Telstra into a strong viable player in the new All Digital Telecomms environment.
        • A major lesson was delivered in 1994-1997 to all Telecomms Service Providers, Regulators and to Policy makers in Government with the commercial failure of the dual Cable TV rollout by Optus and Telstra.
          • This was a signal event and salutary lesson in Natural Monopolies of Infrastructure: each player failed to complete their network, neither player dominated the market for years, between them market penetration was abysmal and within 5 years, both players had written off around half their investment.
          • Yet the Coalition did nothing then and for the rest of its term in office.
          • Continue to ignore these lessons in Market Reality.
  • Cost-Benfit Analyses are meaningless without defining and measuring the Benefits. [U1]
    • This notion came up again on Inside Business over the weekend.
      • When people who understand business & financials and have no political or technological axe to grind, write-off a central Policy Plank, there's a serious issue.
    • How narrowly will the Coalition define "Benefits" in their post-hoc CBA witch-hunt?
      • Whole National Interest model:
        • Business, Government, Essential Services (health/education...), Consumers
      • Telcos, RSP's, NBN Co, Government Revenue & Financial Outcomes.
      • Government Financial outcome only: this will likely be the focus of the Coalition 'Inquiry'.
  • The Coalition modelling is highly biased and woefully incomplete [U1]
    • Whilst very rightly & throughly/reasonably the current NBN Co plan is stress-tested,
    • there is NO "Rivers of Gold", or more optimistic scenarios tested, and
    • NO stress-testing is done of the Coalition plan. Minimally this needs to be done:
      • Same 4 stressors applied and
      • Additional stressors unique to DSL/Node model.
        • Such as Telstra delays and demand higher price for sale of whole asset.
        • Remediation costs run out of hand.
        • DSL node/line costs are over 25% of FTTP.
        • Nodes can't support 
        • Network upgrade to Full Fibre is more than 109% direct to Full Fibre.
  • The Coalition does not mention "upload" or "upstream" rates. [See: TelecomTV.]
    • Impact: Services like Skype, Video chat, Gaming, P2P file sharing and legitimate Dropbox file transfer and backups all depend on upload.
      • More importantly, congestion of the upload kills download speed.
      • WIth TCP/IP, all transfers require low-latency on the return path for the 'ack' packets. Performance of both sides of a link needs to be "sufficient" for services to achieve reasonable throughput.
      • If "daddy does an upload or backup" during the family's favourite TV show, the download will stop. Even if the only download is an 8Mbps video stream over a 25Mbps link, it will stop playing.
      • Alternatively, emailing a video or collection of images/songs to other people will have the same effect as an upload. Though will be much harder to notice and control.
      • As the sites on bufferbloat point out, simple bandwidth arithmetic does not work with TCP/IP. There are very complex effects, with Queues and feedback, operating.
    • Likely Problem: On DSL, download speeds can be traded for upload speeds - the amount of RF spectrum available is fixed and must be allocated per "user profile".
      • The Coalition may be intending to achieve 25Mbps by only allowing 1:10 upload rates, or 25/2.5Mbps, not the Fibre rate of 25/10 and 25/20 Mbps. Worst case is they'll only provide 24/1Mbps, (ADSL2+ Annex A) or 24/3 (ADSL2+ Annex M)..
      • To understand this is not simple arithmetic with DSL rates, i.e. there is NOT a total of 27.5Mbps available to be carved up, look at the symmetrical DSL (SHDSL) rates for sale: 2M/2M or 10 times slower.
  • Telecom TV mentions other deficiencies in the Coalition NBN Plan:
    • The "drop-dead" project risk of VDSL2 and "vectoring" either failing in the field to deliver on performance expectations or proves much more expensive. The difference between "theory and practice" well know to Engineers.
    • the compound impact of competition from HFC & 4G wireless on the revenue side of the project plan.
      • The financial risk is that almost all low-cost FTTN services will be undercut by the competing services, leaving NBN Co only able to sell to high-cost premises ('edge case' FTTP, Fixed Wireless, Satellite) removing all implied subsidies.
      • This pushes per-premise costs up 2-4 times.
    • Problems with the quality of the last 1500m of the copper.
      • Either lowered speeds, much more than 22% Fibre to Premises, Fixed Wireless (more expensive per premise), more remediation or shorter runs and more nodes (higher overhead and maintenance costs).
    • In-depth geomapped assessment of FTTN deployment costs may reveal much higher costs, or Fibre-to-Premises rates, than the UK/European FTTN experience.
    • The failure to address:
      • The economic fragility of the Plan. What's their fallback if the plan is infeasible?
      • Any contingency planning for the technologies that are untried and untested in our environment. Saying "it's worked elsewhere" is NO guarantee it will work here.
    • Impact: All these deficiencies are major or "drop-dead" project risks.
    • Likely Problems: Included briefly above, all severe or catastrophic impact.
  • There is no mention of the full cost of the FTTN network, nor the breakdown of the $20.5 billion CapEx between the four networks to be built.
    • No mention of the install cost and delays of Nodes and ISAM's in exchanges.
      • Planning permission is required for all 60,000 nodes.
      • TransACT in 2002 spent around $20M on just 
    • No direct statement of the per-line cost assumptions. $900/line is used in a single example.
    • Impact: Is the cost of the FTTN $900 * 8M = $7.2 billion or $4-$5 billion?
      • By not stating the cost of the various network components, nobody can challenge their total of $20.4 billion in CapEx.
    • Likely Problems: This is very deliberate decision aimed at hiding their assumptions.
      • Because of the level of minutiae included when criticising the current NBN Co plan, we know that this is a deliberate withholding directed towards some Political advantage.
  • Whilst an indicative figure of "60,000 nodes" is given, no information is given on the architecture of the network nor of the most crucial parameter: the node-premises maximum distance. Nor the number of ISAM's that will be installed in exchanges.
    • Are they planning 400m, 800m or 1500m maximum node-premise distance?
    • Whilst the Coalition admits DSL performance declines with distance, they neglect to mention the most salient fact: it reduces exponentially or worse with distance. There are many good diagrams they could've included to illustrate this.
    • Impact: If the UK FTTN is built to a 400m rule, nothing about it translates to us if we're using a 1500m rule.
      • This is a critical factor and very well known to all players.
      • It's omission is deliberate and for some specific purpose.
  • No criteria are given for "Cost-Effective" allocation of premises to the Fibre Network, nor the process by which individual premises or areas will win the Fibre Lottery.
    • Just who will be the lucky 22% to get Fibre laid on for free?
    • Impact: Will we be surprised to find that Coalition seats will magically be judged as instant winners of the Fibre Lottery?
  • No direct statement of the expected and worst-case payments to Telstra.
    • Telstra shareholders are likely to be very wary of any Coalition promises.
      • The T2 investors that paid $7.40 in 1999 are underwater by 50%, still.
      • The T3 investors that paid $3.60 in 2006 have been underwater for most of that time. They will not be kindly disposed to vague Coalition assurances.
      • Only since 2011 and the NBN Co agreement has the shareprice recovered and then very sharply, up 70+%. Turnbull has trumpeted the Coalition plan is expected to add 2% to the Telstra share price. That's not a deal that would entice me!
      • Abbot/Turnbull need to be offering a very good deal to persuade/entice what are now a very cynical and cautious group of investors, not voters.
      • Unlike election promises, they need to be making hard-cash offers to the owners of Telstra. Without a contractual obligation, it's all hot air and won't sway shareholders.
    • Turnbull is relying on his claim "Telstra has said publicly that the Copper CAN is of near zero value" and has made statements to shareholders, via the Press, that "Telstra won't be affected by the Coalition NBN".
      • Only the Board or the CEO should offer specific guidance to the market.
        • Is this an ASX/ASIC violation? Outside of Politics, it certainly would be.
      • Only the Board and ultimately the Telstra shareholders can make this decision to enter into this contract or not, and at what price.
      • As per the 2010 Telstra shareholder vote, independent detailed financial modelling needs to be done and supplied to shareholders to satisfy due diligence requirements.
    • A figure of $1500 for PSAA payments to Telstra is claimed. The details of the Telstra arrangement are confidential, the only public record details are "around $4 billion in post-tax NPV benefit".
    • Either the Coalition is guessing, in which case they need to show how they arrived at their figures (my estimate is $1200/line and 70% cutover), or Turnbull has just released a previously confidential figure.
      • Neither option covers him in glory.
    • Prudent management principles suggest the project should not even commence detailed planning until this issue is resolved and a contract signed.
    • Impact: Any delay by Telstra shareholders in accepting the Coalition proposal will impact the project start date, already very aggressive and optimistic.
      • Telstra shareholders may decide on a per-line sale prices of $2,000-$2,500, a realistic market value to an FTTN operator, not the $1200-$1500 PSAA price assumed by the Coalition.
      • This additional cost, $8-12 billion, will put the Coalition Plan "underwater", i.e. more expensive than the current full-fibre plan.
    • Likely Problems: At the very least a 12-18 month delay to the project start is probable, eliminating the claimed benefit of "We'll get it done 24 months sooner".
      • If Telstra shareholders demand a reasonable market value for their Copper CAN asset, the Coalition Plan is a non-starter.
  • No statement is made about transferring maintenance costs and staff from Telstra to NBN Co, or contracting them back to Telstra. The only indirect statement is a projection of "$90/line" in maintenance.
    • If Telstra is asked to contract maintenance until 2019, then transfer assets and staff to NBN Co or another contractor, from past experience, they will drive a very hard bargain.
    • The rollout cannot be commenced with the maintenance contracts being in-place.
    • Impact: The OpEx will increase 50-100%, forcing higher wholesale pricing for the Copper service.
    • Likely Problems: Both schedule delays in starting the project and OpEx significantly higher than $90/line.
  • There is no mention of NTD's (Network Termination Device) or who assumes the cost for installation, supply and maintenance. Nor for additional services & NTD's when maximum data rate is reached or the customer requires more UNI ports.
    • This is based on a search of the PDF files and a quick eyeball scan. Because the documents were constructed to obfuscate text, this is not definitive.
    • Specific NBN Co NTD's are required that are functionally identical to the existing Fibre, Wireless and Satellite units.
      • VLAN (802.1qinq) compatibility for ethernet bitstream provision.
      • 2 phone lines
      • 4 UNI ports
      • remote administration and management
      • An are much more expensive than existing ADSL router/modems and require very specific capabilities not found in any existing retail units, to my knowledge.
    • An NTD is necessary for any Internet services.
    • The NTD now becomes the 'Edge' of the Telco network.
      • The subscriber cannot install their own because of this.
      • Only Austel certified cablers are allowed to touch Telco wiring and devices.
      • The least cost of an install paid for by a customer, will be $200 for the NTD and $250 for 1-hour labour to install inside the premise: $450.
    • Impact: Is this $3.6 billion (8M * $450) included in the $20.5 billion CapEx or is it being shuffled on to the unwitting consumer, along with a $3-$5,000 hit to upgrade to Fibre "sometime"?
      • If this is to be paid for by the consumer, why are they singled out when Fibre, Wireless and Satellite service will have NTD's supplied and installed?
      • Is the real comparative CapEx budget $20.5 + $3.6 billion = $24 billion?
  • Are all NTD's and install costs going to be passed in full onto consumers?
    • One resolution of the asymmetry of only charging Copper service customers for NTD's is to charge everyone for them.
    • Impact: The Coalition needs to clarify both the cost & payee for Copper NTD's and if they will be forcing these costs onto all consumers/
  • ADSL Disruption and transfer from existing DSLAM provider.
    • "no disruption" only applies to PSTN (phone) services.
    • ADSL services are terminated to every premise served by a node when the cabling is connected to it. This is because there is no longer a copper path back to the ISP's DSLAM.
    • As noted above, specific NTD's are required to connect to the NBN. Existing ADSL router/modems are not suitable.
    • This forced resumption of a commercial asset, the working ISP DSLAMs, will cause actions for compensation.
      • I've not noticed in the Coalition documents any amounts to compensate the many ISPs for the loss of economic assets.
    • Impact: The Coalition claims in its major graphic "No Disruption".
      • This is only true of phone services.
      • The concurrent installation of a Fibre service is not only possible, it will be the standard. It means the service can be installed and tested before being handover to the customer. A software transfer of both phone and internet services will make this truly transparent to the subscriber.
  • Data Volume Demand Growth model: continuing until 2028 well above 50%, let alone 30% (current 63% ABS). Compare to this spreadsheet. [referenced below as well].
    • The Coalition does not outline either:
      • the Data Volume Growth curve they used in their modelling, or
      • the capacity of their FTTN. Most importantly:
        • the number of lines per node,
        • the capacity of the Node uplinks and
        • the Node "backplane bandwidth".
      • The Alcatel 7330 might be a suitable, cheap platform, but won't work for us:
        • It only has a 7Gbps backplane.
          • While it might support 200 lines, it only supports 35Mbps/line.
          • You might pay your $5,000 for a Fibre upgrade, only to find the node backplane is maxxed out and you get much less than 100Mbps.
        • It only supports 1Gbps uplinks.
          • Or 5Mbps per line in a fully configured node.
    • Impact: The Coalition needs to fully disclose its assumptions and design to allow the electorate to properly assess its policy.
      • This is a "drop-dead" for consumers.
      • Cheap, low-performance ISAM's will make the service unusable for all consumers on a node. This will be very hard and very expensive to recitify.
      • The Gungahlin experiment, where subscribers had exactly these problems with Telstra RIM's/CMUX's, informs us that once a Telco has your money, you can "go whistle" for any service upgrade.
  • Remediation costs, including removing "Pair Gain" systems, loading coils and "bridge taps" as mentioned in the Telstra 2005 technology briefing as a major work issue.
    • Trujillo had noted that from 2001-2005 Telstra had underinvested in CapEx and OpEx (e.g. copper line maintenance) at least $2-$3 billion. Following the massive technical layoffs at Telstra, this situation will have worsened.
    • In responding to the FANOC proposal of 2008, Telstra said it could rollout an FTTN with 40% few nodes by re-laying the copper CAN optimised around Nodes, not Exchanges.
      • This is a major planning exercise that must be carried out in fine detail.
      • The budget and schedule allowance for this critical step is missing.
    • There has been no definitive audit of the state of the Telstra Copper CAN for DSL use, especially the last 1500m, with assessment of remediation needed for a stable, long-term DSL service. The NBN Co assessment has been for retasking ducts/pipes for Fibre.
      • This is a project-risk of the very highest order.
      • Examining in detail 20M copper services over 150,000km "road distance" will be a major and expensive undertaking in itself. More so if useful electronic measurements are taken from the pillar to the customer connection point: this cannot be automated as there is only remote testing equipment installed in exchanges. The last 1500m has to be measured and that AFAIK, has to be done by hand.
      • The project-risks are three-fold:
        • Delay in starting. The whole copper network in each PoI region needs to be assessed/audited before a financial decision can be taken on whether an FTTN is cheaper at all, or just which areas will be Fibre and which Copper.
        • Project schedule over-run due to Remediation times in excess of 200% forecasts or Fibre required over 40, not the budgeted 25%.
        • Project budget impact through delays in Copper rollout, Remediation budget overrun and over 50% Fibre budget over-run and implied schedule over-run.
    • Impact: This is a "drop-dead" project-risk.
      • Both in pushing back the start date while proper assessments/surveys are made, and
      • in forcing large Budget over-runs.
    • This is a misleading statement, wholly untrue in the way it is used:
      • "Using the more realistic assumption that usage increases by 30% per annum, the CAGR for the basket of AVC and CVC services purchased by end-users is 19%". Lifting the inflation-adjusted wholesale price of a basket of plans from $28/mth to $145mth. Source: page 13 of Coalition "Background Papers".  
      • This spreadsheet provides the data for the relevant scenarios of Data Volume demand Growth.
      • The real deal is NBN Co will be charging 4 times less for services in 2028 under this scenario. Telstra wanted further discounts.
      • There's a secondary effect:
        • Consumers are NOT obliged to pay this amount.
        • Only consumers who followed the 30% CAGR trend and who find utility in the service, will pay the $145/month wholesale price.
        • Just like now with Internet and Mobile Phone plans, Customers will choose their own price-point based on their own preferences.
      • An effect not covered by the SAU or in the Telstra submission to the ACCC was the price of backhaul capacity: unless it falls by at least the same rate, those costs will come to dominate the Retail cost of Plans.
        • CVC pricing, the price/bandwidth at the PoI, is covered by the SAU and while it will drop many times faster than inflation under the 19% CAGR scenario, NBN Co does not control the cost of backhaul that RSP's.
        • Telstra, as a major wholesale provider of backhaul and owner of the most extensive internal backhaul network, has both a significant cost advantage over RSP's and can set backhaul pricing as it wants.
        • If Telstra and other backhaul providers only halve their pricing, the cost to RSP's for the 30% CAGR basket doubles.
      • All the Telcos know that these huge CAGR's are a River of Gold and they want to cash in on them in as many ways as they can. Optus and Telstra were both trying to pitch to the ACCC for a greater share of the customer dollar, at the expense of NBN Co.
      • The Coalition is criticising NBN Co not for cost increases but for creating and allowing a fantastically profitable service for Telcos and one that consumers find so compelling that they are willing, like their smartphones, to pay an increasing proportion of their disposable income towards.
      • Impact: The Coalition is deliberately spinning the facts. Once the electorate understands the depth of these misleading statements, they won't be pleased.

    No comments:

    Post a Comment

    Note: only a member of this blog may post a comment.