Monday, 17 June 2013

"NBN or SBS?" What Barnaby Joyce said about himself

I happened to catch Barnaby Joyce flip the tables on Ellen Fanning by saying:
I'll start with one, do you believe the SBS is more important than the NBN?
An innocent & fair question? No, not nearly.
While it sounded wrong to me at the time, it's taken me two weeks to be able to articulate just what was so wrong...


Firstly, you have to love the way Joyce took control of the interview and the agenda. Masterful use of the naive, humble-bumbling-country-boy persona he's deliberately manufactured.

Secondly, not only is this a question based on a false premise, that there could be an either/or choice, it's a question underpinned by Accounting questions, Barnaby's field of professional expertise.

He's either believes what he's saying, in which case he's incompetent in Accounting, or he's dissembling and misleading and knows exactly the argument he's making. Nothing in-between.

I think we can presume because of his years running a successful practice, Mr Joyce is actually highly competent in Accounting and knows exactly the false premise and fallacies embodied in his question.

This was over-the-top disingenuity, even monumental hypocrisy.

Here are the three falsehoods in that "innocent" question, seemingly chosen at random:
  • The expenditures cannot be compared: one is recurrent (yearly) to run an enterprise, while the other (NBN) is a one-off project, building an asset.
    • In for-profit enterprises, SBS would appear in the Profit and Loss statement, and investment in the NBN project would appear in the Balance Sheet (and Cash Flow Statement as Investment activities, not operations, nor financing.)
  • The NBN is an investment, undertaken because it has a projected Return on Investment (of ~7%).
    • The NBN will, if things go to plan, more than pay for itself, the Government is running it as a for-profit business, not as "public good",
    • whereas SBS is not a business, thought it does raise some of its own operating income.
    • SBS (and the ABC) are funded by the Government solely because it's a good thing and its have a positive social value.
  • SBS is "on-book" and NBN Co is funded off-book, precisely because it's an investment that is expected to return its principal and at least 7% on top of that.
    • The money borrowed by the Government to fund NBN Co, around $30 billion - the rest will be Debt that NBN Co has to raise itself, will be paid back, and more. That's what investments do.
    • As much as we love and cherish SBS and the ABC, they aren't expected to pay their own way or make a return, just like pensions and welfare benefits, we as taxpayers and voters put money towards these activities because they are worthwhile social activities, not investments.
Barnaby knows all these arguments and their implications only too well. He blind-sided Ellen Fanning and left her floundering because these questions are arcane.

What we saw when Joyce very quietly & effectively took control of the interview from one of our best, most experienced journos was a tour de force by a very competent, coldly calculating politician. This display shows up both Warren Truss, current leader of the Nationals and Tony Abbott as the very mediocre performers they are in debates and interviews.

Barnaby, "the gift that keeps on giving" as he self-describes, is not some bumbling fool from the Bush, but a consummate politician and debater, showing himself to be more than a match for one of our elite journalists. Is he a latter-day Joh Bjelke-Petersen, smart-as-a-whip masquerading as a fool? Better to have people under-estimate you.

This question and line of conversation can't have been accidental - people with this high-level of debating trade-craft prepare, carefully and completely. We probably didn't get to see half-a-dozen other pre-prepared plays.

That Joyce repeated in a line he knows to be demonstrably false, the NBN will cost "$90 billion", not the reality "the 1 in million worst-case scenario is the NBN would cost $94 billion, if nothing were done by the project managers in response to project stressors".

Making this statement shows that Barnaby was deliberately disingenuous in his statements and monumentally hypocritical in his conflating of two completely different, incomparable types of expenditure.

I won't go near the assertion that "all government debt is bad" and that we've got massive debt that is ruinous... Portugal, Ireland, Greece, Spain, Italy have massive debt and consequential catastrophic unemployment rates and social problems. Japan has been stuck in stagflation or debt-deflation for two decades now - they've got real economic problems. The UK and the USA have economic problems that, to quote Letterman, "I wouldn't give to a monkey on a rock".

Turn on the TV and watch some international news to understand just how well off we are...

Finally, there is a real comparison question that would illustrate Barnaby's point, on the "Sophie's Choice" nature of politics:

Do you believe the SBS is more important than the AIS?

That is an Apples-and-Apples comparison and would illustrate to many people just how hard the multi-layered beast politics really is.

But Barnaby wouldn't go there, he couldn't make gratuitous attacks on the Government, nor "be on-message". The AIS, even with the recent failure of the swim team, just isn't news.

Saturday, 15 June 2013

NBN:Policy Questions for the Coalition

With the 43rd Parliament about to enter its final sitting weeks, it seemed a good time to recap NBN Policy Questions.

At first blush, the Coalition NBN Policy cannot meet its claimed savings and Whole-of-Project costs must be higher than the current direct Fibre plan. There are many important questions unanswered.
  • Will the Coalition be offering, for VDSL2/FTTN customers, a pure-digital Network Termination Device (NTD) with 4 UNI-D ports identical to the other two transmission networks: GPON/direct Fibre and Fixed Wireless? (Current Satellite NTD's have a single LAN/UNI-D port)
    • Will VDSL2 NTD bulk installs be offered to customers in the roll-out phase, the same as default Fibre PCD installs, or made available, like Fibre, by appointment?
    • This is an important detail:
      • Will identical offers be made to customers across the four transmission networks?
  • Will the Coalition be providing an independent, expert review of its NBN Policy costings well before the election?
    • The claimed $17 billion savings seem impossible when the advertised difference is, at best, a savings of $450 each for 9M premises, or a total of $4 billion.
  • The presumed single access-charge and lower average line access rates of VDSL2 FTTN, plus absence of high-end usage, must translate to significantly lower revenues for NBN Co, both in Access Charges (AVC's) and download volume (CVC's).
    • Any review of the Coalition NBN Policy must include a forecast of this foregone revenue.
    • In operation, NBN Co will be able to track this accurately, from the 25% of fixed-lines on direct-Fibre. Will the Coalition commit to regular release of these broken-out figures?
    • Will the Coalition publish estimates of lost revenue from "Cherry Picking", foregone revenue from the HFC area not being overbuilt?
  • Whilst the Coalition ran four stress-tests and one worst-case scenario against the 2012 NBN Co Corporate Plan, the following studies are needed for an Apples-and-Apples comparison of the two Plans:
    • Reworking NBN Co stress-tests with current, valid figures, not wildly inflated as current.
    • Running the same, or equivalent, stress-tests on the Coalition NBN Plan.
    • Running two high-growth scenarios on both the current NB Co and Coalition NBN Plans, with:
      • 50% CAGR (Compound Annual Growth Rate) in download volume demand, seen in the ABS figures.
      • 66% CAGR forecast in the CISCO VNI report for Australia.
  • The VDSL2/FTTN proposal includes significant additional costs transferred onto householders. Any review of the Coalition NBN Policy needs to include an estimated Whole-of-Project cost, NBN Co costs + householder costs to first data connection, to allow an Apples-and-Apples comparison:
    • Adam Internet, the only VDSL2 service in Australia, charges over $150 for VDSL2 modems. Because they offer a pure-digital VDSL2 service, no line-filters or splitters are needed and customers can install the modem themselves, unlike the hybrid telephone-digital service the Coalition will offer.
    • Telstra charges $300 or more for service connections that require a Technician visit, necessary if a Central Splitter is required.
      • If both are needed, it easily cancels all anticipated savings.
  • NBN Co may well mandate Central Splitters for the VDSL2/FTTN to meet service quality and network reliability targets.
    • If so, who will pay for the certified technicians to install the Central Splitters on the Telco side of the customer boundary, the customer or NBN Co?
    • If customers later have the option of upgrading to a pure-digital VDSL2 NTD, then:
      • Will they get a rebate for the install of the Central Splitter, if they've paid?
      • If Splitters are required to be removed, who will bear that extra cost?
  • The Coalition readily admits that a VDSL2 FTTN is only a temporary network. It has never said how long an FTTN will need to operate to recover its costs, nor how long their plan estimates it will operate before being replaced. Mr Turnbull has used example times of 10 and 15 years when explaining Net Present Value, but never revealed a figure. This raises two critical issues:
    • How long does the Coalition NBN Plan assume before the VDSL2 FTTN is replaced?
    • Who will pay for the replacement of the FTTN?
      • Will it be another Government-funded infrastructure project, which the Coalition have already stated is anathema to them?
      • Will it be funded by NBN Co out of revenue or by raising Debt?
      • Will this be the trigger to privatise NBN Co?
  • The Coalition NBN Policy sets a hard upper-limit on funding to NBN Co of $30 billion. What happens when, not if, NBN Co exceeds this figure because the Coalition costings are so wrong?
    • Will a Coalition Government force them to stop their roll-out work or declare them bankrupt?
    • Will the Coalition allow them to raise Debt, as Bonds, on the open market?
    • Will the Coalition privatise NBN Co at this point, offering shares first to their major creditors, Telstra and Optus?
  • Under any of these FTTN termination scenarios, what happens to subscribers and what they'll be charged:
    • For the forced replacement connection.
    • For access and download volumes, AVC & CVC wholesale charges to retailers.

Thursday, 13 June 2013

NBN for Real People: Networking

Part 1 in a series. My take on "Computer or Digital Networking":
(compare to the New Oxford Dictionary definition at the end)
Digital or electronic devices talking to one another, often to facilitate conversations or messages between people.
Almost everybody is familiar with, and uses everyday, electrical and communication networks:
  • Electricity, also called "Power Supply"
  • Telephone
  • Mobile Phone
We have many other "networks" that we rely on in daily life:
gas, road, rail, water, sewerage, storm water, bus, train, tram, airline, TV and radio, food & grocery delivery, ...
Networks are generally drawn, not described in words, showing their relationship to a branch of Mathematics called Graph Theory.

Network drawings, or diagrams, have two essential features:
things and connections between those things. Connections are drawn by convention as lines, and many sorts of symbols are used for the "things" depicted, especially to distinguish and group them. Colour and other perceptual cues are used to make these diagrams more easily understood.
Importantly, the Network Diagram does not have to be "drawn to scale" like a map to be useful, often the reverse. One of the most famous Network Diagrams, is the Metro Map. It's useful precisely because it's not an accurate, but highly stylised, representation of the Network.

Two types of Networks

Did you notice both "rail network" and "train network" in the examples? Ditto for road and bus?

There's an important difference there, infrastructure, and services using them, can both form networks:
  • Road and rail networks are fixed, or static. They're networks of physical things, often called "infrastructure", they don't move ("static"), can support multiple uses of them (passengers, freight, large & small vehicles, scheduled or on-demand) by multiple, different users or organisations. They are only of indirect value to people. 
  • Bus and train networks are not tangible, physical objects, but are dynamic and provide services directly of value (or "utility") to people and organisations using physical infrastructure and "equipment" like rolling stock, buses, trucks, cars, ...
  • Services, such as ferry, shipping and airline networks, can operate without a tangible, built infrastructure. Shipping lanes and aviation routes are lines drawn on maps, not physical entities, though they always end with a port or airport, to allow passengers & freight to load/unload.
An example of an Infrastructure, or physical, Network:

Think of your electricity power supply and how it gets to you from the power stations.

There are lots of wires & switches, transformers and branching of the wires, until a wire runs past your house, a wire runs off that to your house's fuse box & meter + main switch, and then more wires run around your house, ending in power-points (power outlets) or fixed appliances or lights + switches.

These power outlets allow you to access electricity to power devices you want to use as you want to.

That's the Electricity network.

The service is the being able to draw power on demand in your home/premise, which is different to the physical network (infrastructure) of wires, switches, transformers and generators that merely provide access to that power.

You buy electricity from a Retailer, or in NBN-speak, a "Retail Service Provider" (RSP). These are the same people you now know as ISP's (Internet Service Providers) or full "Telco"s like Telstra and Optus, short for Telecommunications Company. Telcos providing a full range of services across multiple types of networks, such as telephone (landline), mobile phone, leased lines and Internet services.

Services

Before you can have a service, you need a physical connection to the wires.

Before the Retailer turns on your service, they have to get money from you, create an account and then activate your service. To do that, your Retailer has to deal with the people it buys services from, in NBN-speak either a wholesale provider or aggregator that act as a go-between between small Retailers and wholesalers.

The current main wholesaler, NBN Co, brings together lines for 100,000 premises at a "Point of Interconnect" (PoI). There are 121 PoI's. On one side are the wires to your house, on the other, each RSP has to provide its own connection back to its offices or data centre. Another set of companies own and resell these links, or backhaul. In the Electricity world, wires to houses are called the Distribution Network and  long-distance, high-voltage lines are called the Transmission Network which finally connects to the various Power Generators.

With the Internet, RSP's connect between their various facilities, "up-stream" to each other via either "peering networks", "Tier-1 providers" or directly overseas via international links owned by "Carriers" or Telcos.

But before you can connect to a network, it has to pass your house or premises.

These steps create three counts to measure the size of the Network:
  • premises passed
  • premises connected
  • premises active (or activated)
The "take-up" rate is the proportion of premises activated to premises connected. [believed to be a correct definition]

Connecting to, or using, a network.

Connecting to the Electricity Network, at home, is as simple as plugging a toaster/jug in.
A plug, on the end of a cord or electrical cable is inserted into the socket of a power outlet.

Connecting to the Internet at home, once you have an active connection, is the same.

Either plug a network, or ethernet, cable (usually blue) into a socket on a modem, switch or router, or connect wirelessly, often called "WiFi" to distinguish it from wireless mobile phone (also called "3G" and "4G") and wireless "bluetooth", a short range way to connect devices like headsets or hands-free phones. You might have used "bluetooth" in your car with your mobile phone or smartphone.

You hire certified professionals to rework the wiring in your house and your Electricity connection, add the fuse box and attach your house to the main supply. Power flows out of the power stations, across the main transmission lines, into local distribution networks (plural) and then into your home for you to 'draw down' when you need/want.

Similarly, larger enterprises hire trained networking professionals to look after their internal Data Networks. Householders and small businesses can hire professionals or attempt to do the work themselves.

Communications Networks I

There are many devices that need to be configured to make the internet function in your house. None of those normally matter to you, or should be seen by you. It's the job of the professionals working for your Retailer to provide you with a simple, reliable and understandable connection, just like the Electricity Network.

There is a major difference in how you'll connect to the NBN and how you buy ADSL broadband services over your copper phone line now:
  • currently, all an ISP provides you is configuring an activating their end,
    • You have to provide and configure an ADSL modem/router yourself, trust a friend to do it or, if you're very lucky, some ISP's will sell you a pre-configured modem/router that you get to plug into the phone line and computers. A very few companies sell professional in-home support to make it work.
  • in the new world of the NBN, NBN Co will install a device, an NTD (Network Termination Device) in your home that your RSP can configure and administer. It brings a working Internet connection into your home, without you need to buy or configure anything.
    • You can plug a single device directly into your NTD, but it's unlikely you'll do just that, to get the best use out of your Internet connection, you'll want to spend more.
    • At the very least, for security reasons, you should install a router/firewall. These are often available with WiFi, USB and multiple ethernet ports, useful in home offices.
    • Most people will choose to wire their houses, just as Enterprises wire their offices. This means adding a router/firewall for security, ethernet switch(es) to connect more devices, a WiFi gateway to use laptops, tablets etc, and having ethernet sockets installed in wallplates.
    • Fixed wiring & wallplates have to be installed by certified cablers.
    • Anyone can plug in a lead (non-fixed cable), a switch or router.

Communications Networks II


A (digital) communications network is pretty much the same as the Electricity Network. Think of the phone network, a bunch of wires and specialist equipment connecting electronic devices together.

The only difference is there's more two-way flow:
while you download/draw-down lots of bits, you also upload a lot.
Data doesn't now come to you in a continuous flow, but in packets. In Olden Times, data only came via  a continuous stream: exactly the same transition as Direct Current (DC) to Alternating Current (AC).

Batteries use and produce DC - why you have special devices (chargers) to change AC into DC. It's
good for local use, bad because it doesn't work over long distances.

Resistance in the wires creates heat and the power is dissipated along the way, until nothing usable is left. Thinner wires create more heat, thicker wires less heat, but get very expensive. Resistance heating is how electric heaters and old light globes (incandescent) work.

AC has a bunch of useful properties that make it possible to transmit power efficiently over long distance, all related to the stop-start nature of it. There are still losses, but they are managed and cabling costs can be minimised because the voltage can be changed easily.

Phones and the Internet

A major difference between electricity and data is identifying destinations. Electricity is 'anonymous', you can't identify where any piece of power came from, and every watt is identical (though voltage
may vary).

But data is like a phone call, every connection needs two ends or parties (at least) AND conversations are 'directed', both parties need to be exactly identified AND the bits are all different, and order is important.

Think of phone calls - you want the sounds reproduced exactly as they come out of your mouth. If they are jumbled up, it's not speech anymore, the principle that scramblers use.

Something less obvious with data is that every download you make has to be checked for errors. Each of those data 'packets' is checked when it arrives and a tiny message sent back to the sender saying "OK" or "Not OK, send it again".

 These are "Acknowledgement Packets", shortened to "Acks" and "Nacks", or Negative Acknowledgements.

For most applications, you want a perfect copy of the original data and are willing to trade time (resending) for perfection. A spreadsheet, email, digital photo or movie won't be intelligible if the bits are corrupted


"Download" capacity is as important as "upload" capacity. Data & phone networks need to work in both directions. If the Ack packets can't get back to the other end, it will slow sending, or stop all together.

Speed, Latency and Throughput: baffling you with B/S

There is no one measure of the "speed" of your Internet connection.
As an informed consumer, you need to understand the difference between three measures, how they relate and the questions to ask your Retail Provider or ISP when "the Internet is Slow".
  • Politicians and salesmen will emphasise the raw speed of the line into your house. It may be 12Mbps (Mega-bits per second, not bytes, they are eight bits long) or 100Mbps.
    • Obviously, 100Mbps is always faster than 12Mbps! Not necessarily. Which is why NBN Co goes to some lengths to try to inform consumers of what they can expect. 
  • Congestion, Throughput and Latency are networking terms you need to become familiar with.
    • Anyone who's commuted in a large city is already too familiar with the concepts, if not the terms:
      • Congestion, or Traffic Jams, as seen on roads in Peak Hour.
      • Throughput is how many cars an hour travel down a road, or for an individual car, the average speed for the entire trip.
      • Latency is the time taken for a car to get from Point A to Point B.
        • Round-trip time, From A to B and back again, is also important on the Internet. If A to B is fast, but B to A is congested, connections will slow down.
      • Link Speed, also Access Rate, is the fastest speed you travel at any point on your journey, not the much more important average speed. If you spend 15 minutes travelling at 100kph on a freeway, and a total for 45 minutes each end travelling just 5km, which matters most to you, the door-to-door time or whizzing along the freeway?
  • If you buy a 100Mbps link from your Retail Provider, your RSP/ISP may only purchase 2.5Mbps of backhaul capacity for you, a 40:1 "over-subscription".
    • Most of the day you'll not notice you're sharing the link with 39 others!
    • At 6:30PM when everybody sits down to watch TV, you will all notice and not be happy.
    • Although you have 100Mbps inside the NBN Co network, you will only get 2-4Mbps throughput from your RSP.
    • Retailers offering cheaper plan can save costs outside the NBN Co network, they can buy very small "backhaul" and "upstream links".
    • A 12Mbps link without congestion inside the provider network will beat the pants off a 100Mbps connection during Peak Times, but not for the rest of the day when its quiet and you, like everyone else, aren't using it.
  • Congestion leads to high-latency which can lead to problems with streaming audio and TV, downloading files and webpages and especially, making phone calls over the Internet.
    • Congestion inside your RSP will make any real-time service, like phone calls, Skype or streaming video, unusable during Peak Times.
    • What you're interested in as a consumer is the sustained throughput that your Retail Provider is prepared to guarantee.
  • ADSL, VDSL and Fibre to the Node, FTTN, networks have a common problem: congestion on the Fibre uplinks from the copper ADSL/VDSL modems or DSLAMs (Digital Subscriber Line Access Multiplexers) back into the NBN Co network.
    • Direct Fibre to the home doesn't have this problem: you have a guaranteed access rate right back to the end of the NBN Co network, the PoI.
    • Fixing congested uplinks from nodes is very expensive and very slow. This was a long-standing problem with Telstra RIM's, a hybrid phone & broadband node network, in Gungahlin, that took many years to fix.
    • It is not considered "Cost Effective" by Telcos to install extra-capacity in the nodes or the fibre links that connect them back to the network. Demand for Internet capacity has consistently outstripped forecasts, leading to persistent local congestion for most users of FTTN services.



Conversations in Real-Time are different to downloading files and webpages

Like CD's, audio, such as a phone conversation, can be "lossy" - our hearing can tolerate some missing data, BUT it can't tolerate delayed or scrambled data. Because speech is "real time" and connections must work within very tight timeframes, shorter than the "round-trip delay" between two ends, we  can't Ack/Nack & resend.

Especially, our hearing can't tolerate packets taking different times to arrive. If you need 50 packets a
second for a phone call, one every 20 milliseconds (msec), then you can't send nothing for half-a-second and send 25 packets all together in a single burst.

For a phone to reconstruct reasonable quality audio, you need packets to arrive predictably, with little variation. I.e one packet every 20 msec with maybe 0.5 - 1.0 msec variation... We call this "jitter" because when you look at traces of it, they dance around and look "jittery".

Recap

To recap, plugging a computer into a network is identical to plugging in your shaver, toaster, jug or a charger into electricity.

It's the job of Professionals to hide all the complexities from you, the end-user.

The NBN, with fully-installed and remotely configured and managed NTD's is a bit step forward: householder can just plug into

For security reasons and to get WiFi for laptops and tablets, you'll want to at least install a router/firewall. Some people will run fixed cabling around their homes so they can plug-in many devices and connect them to the Internet.

"Download" capacity is as important as "upload" capacity. Data & phone networks need to work in both directions.

"Speed" isn't simply how fast your first link runs, but the sustained throughput, especially at Peak Usage times and the "latency" or delay on the line.

Downloading files and webpages requires no congestion of uploads, or the packet Acks will slow and stop, stalling your downloads.

For real-time services, like phone calls & Skype over the Internet, we can tolerate the odd wrong or lost packet, but can't tolerate not getting packets On-Time, or "jittery"




From the New Oxford Dictionary.

Network:

1. an arrangement of intersecting horizontal and vertical lines.
  • a complex system of roads, railroads, or other transportation routes : a network of railroads.

2. a group or system of interconnected people or things : a trade network.
  • a group of people who exchange information, contacts, and experience for professional or social purposes : a support network.
  • a group of broadcasting stations that connect for the simultaneous broadcast of a program : the introduction of a second TV network | [as adj. ] network television.
  • a number of interconnected computers, machines, or operations : specialized computers that manage multiple outside connections to a network | a local cellular phone network.
  • a system of connected electrical conductors.

Monday, 10 June 2013

NBN: Why Fibre? What it does better than Copper.

Bits are agnostic, they don't care what medium they travel over: wireless of any kind, HFC, Fibre, ADSL, VDSL, ethernet, infrared, power lines or dial-up modem.

One of the arguments against a direct Fibre to the Premises network is "anything you can do, I can do better, or at least as well", or, restated, if people want 25Mbps and VDSL gives it to them and its cheaper, then why wouldn't they go for the "sooner, cheaper and more affordable option"?

An FTTP means:
  • immunity from Thunderstorms and electricity leakage Less damage and danger.
  • higher bandwidth per link (forty times), with less excess capacity required to be installed, subscribers won't be denied service from fully utilised infrastructure:
    • subscriber don't require extra premise connections for increased bandwidth or multiple connections.
    • running out of pairs on the local loop and ports in DSLAMs is common with the copper network. This is extremely expensive to address and often never happens. This is a well-known problem with the copper Customer Access Network
  • predictable and guaranteed speeds: if you're connected, you can get the full range of services.
    • This is critical for employer or agency supplied networks, e.g. Schools and University.
  • guaranteed upload speeds
  • guaranteed low latency and network/traffic prioritisation for Real Time services, such as Telephony and high-quality audio or teleconferencing.
  • lower congestion without peak-hour latency "traffic jams" due to larger FSAM's with upgradeable uplinks. We know from the Gungahlin Experiment that nodes are particularly susceptible to under-diminsioned uplinks or saturated backplanes, both of which wi
  • in-place upgrades to higher speeds with newer technology
and
  • lower maintenance and no major upgrade hurdles.
Guaranteed link speeds and uniform subscriber equipment means:
  • multicast, necessary for the much cheaper & more efficient broadcast services, works as designed using the NBN Co infrastructure directly.
  • A single wholesale pure-digital telephone system, eliminating inefficient & costly equipment and systems duplication.
    • This will remove a large deadweight cost from Telephone charges.
    • It will allow new, affordable and innovative Telephony products to be sold to businesses and premium phone users:
      • internal PABX's won't be needed. All switching can be done more cheaply via the NBN.
      • lower call and trunk costs.
      • seamless integration of multi-medium calls, especially from a smartphone: You can start a call from home via WiFi. move to the car using 3G/4G, then to a nanocell hotspot, back to 3G/4G, then to the office with WiFi, seamlessly and reliably.
  • Off-peak mega-speed plans become possible for home, SOHO and SME backups.
    • Backing up servers at 400Mbps (160GB per hour) allows Terrabyte drives to be copied or backed up cheaply & easily overnight without impacting normal operations of the customer network. This is at the lower-end of current demand.
    • Compare this to the current Telstra rate-card for 10Mbps (forty times slower) symmetrical services: $7,931/month for 'unlimited' usage or just 2.5TB, maximum. 100GB/mth is $1,870/mth with $8 per GB after that. 850GB per month is the break-over point to 'unlimited'.
    • versus ~$100/month for 40Mbps and 1TB currently with ISP NBN services.
    • The NBN Co plan is to halve their download volume charges every 3 years, hopefully this will be reflected in ISP/RSP plans. Things will only get better from here, especially for micro and small businesses.
On top of that, the "Bring Your own Device" Coalition policy has very significant downsides for the service providers. It is the worst possible engineering design.

For the Network Operator, NBN Co, and Retail Providers, ISPs/RSPs, a pure-digital network with uniform Network Termination Devices (NTDs) under their full control is essential for properly running their network:
  • Remote end-to-end line testing is possible simply and transparently.
  • Firmware upgrades can be done securely, efficiently and frequently. This is the minimum necessary for a high reliability and secure, dependable network.
  • If all premises have the same networking capabilities, 2-phone and 4-data, the Provisioning and Operations systems are much simpler and less error-prone.
    • From the Telstra HFC network, we know that software control systems can and do seriously degrade and become dysfunctional.
  • Remote monitoring, control, supervision and administration/provisioning of all network devices, especially end-points (NTDs in Customer Premises) is the minimum now required to run large, complex digital networks.
  • IPv6 (version 6 of the IP protocols) can be uniformly & transparently implemented. This will soon become a necessary migration.
  • Automatic response against a concerted Network attack is only possible with a single, known network infrastructure.
    • An "Internet Kill Switch" can be simply and effectively implemented, allowing rapid response to intentional or accidental network attacks or Denial of Service.
    • individual NTDs, or groups of them, can be monitored and controlled in real-time to stop, slow or mitigate the propagation of Worms and malware.
    • Post-event replay, necessary for detailed analysis, is possible by collecting NTD log files.

Sunday, 9 June 2013

NBN: Unanswered questions on Coalition NBN Policy

Summary:

It is more than likely that the Coalition cannot meet any of its headline promises, "Cheaper, Sooner, More Affordable" and most probable that they know this.

  • The outrageous, over-the-top claims of the LNP from their stress-testing of the NBN Co rollouts have never been tested or reworked by independent, credible experts.
    • The media keeps citing the LNP manufactured figures without questioning them.
    • The Government, who has the resources and in-confidence data needed, has not seen fit to provide credible figures, leaving me to wonder where they stand.
    • LNP members keep espousing what should be completely discredited figures of "$90 billion" and "$100 billion" for the worst-case NBN Co project costs.
  • The NBN Co 2012 Corporate Plan is very conservative about growth and income forecasts. Every release of new NBN Co data has confirmed that actual take-up and consumer download volumes are running far ahead of the Corporate Plan.
    • The ALP has not highlighted this nor allowed NBN Co to release any high-growth scenarios to counter the "stress tests" on which the LNP Policy rests.
    •  The ABS reports download volume growth rates of 50% per year and the CISCO VNI forecasts are 66% per year for the next five years.
      • The NBN Co plan assumes only 30% per year growth in download volume, while reducing volume charges by 19% per year, encouraging consumers to use more.
    • If either of these credible growth rates is realised, and the one constant with the Internet over 25 years has been surprisingly high, sustained growth, then the economics of the current NBN Co project will be very different, another point ignored by the ALP:
      • Break-even will be considerably earlier than 2033.
      • It won't take much for the NBN Co project to become self-funding for full Fibre.
      • With greatly increased revenues, if even current forecast growth rates eventuate, NBN Co will become much  more competitive than current ADSL technologies encouraging higher & faster take-up rates and NBN Co can increase the rate of rollout and further increase revenue growth.
      • On credible forecasts, NBN Co will be returning Rivers of Gold to the Federal Government by 2021.
        • The LNP, obviously, will play down & ignore the prospects of their opponents.
        • But why isn't the ALP even talking about, let alone spruiking, the upside of their NBN Plan?
  • The claimed $17 billion CapEx savings require the LNP plan needs to build each VDSL service for -$540 (negative build costs) or for only one-third of usual estimates, $310, and pay Telstra nothing. The latest estimates by Telcos have been nearer $1,000 per service, fully deployed.
    • On top of the CapEx savings, the LNP is claiming they can operate a much older, more maintenance intensive network for 20-25% less.
  • The LNP plan requires timely co-operation by Telstra, never seen before. Yet there are no Risk Assessments or mitigation strategies in their plan to address this "Drop-Dead" project risk.
    • When Turnbull is questioned on the topic he is tight-lipped, offering only the cryptic and unenlightening response, "They will". He has never offered more than this, confirming that something much deeper is going on.
    • Despite their pivotal role in the project and the commercial ramifications flowing from a probable about-face in Government Policy in just fourteen weeks, Telstra has made no statement to the ASX on the LNP plan.
  • Using the $900 per VDSL service implied, but not actually stated, by the LNP, their CapEx can only save 10% of the build cost, ~$4 billion, for a service they admit is inferior and has limited life of 10-15 years, close to the current construction period.
    • Even if the Telstra costs are moved entirely to OpEx by long-term leasing, the 2021 project costs can be no more than 5-10% direct to the Federal Government.
  • The LNP Plan doesn't include additional householder out-of-pocket expenses over what any other NB NCo customer, Fibre, Wireless or Satellite, will pay.
    • This could mean that all customers will be charged a significant connection fee, a major policy point.
    • Correct project cost estimates would explicitly list hidden costs shifted to customers and include them in a Total Project cost, to allow true "Apples-and-Apples" comparisons:
      • The LNP seems to be claiming $450 per service savings for VDSL,
      • but additional Customer expenses will be more than this, given what Telcos are already charging. 
    • Turnbull has claimed on air that VDSL2 modems will "cost $50", whereas the only ISP offering VDSL2 charges $160 on a pure-digital service.
      • The LNP plan includes a "no disruption" promise, meaning they cannot offer pure-digital VDSL service like the other three NBN Co Customer Access Networks, but a hybrid phone/DSL service exactly as is provided now.
      • NBN Co may mandate the use of Central-splitters in such a hybrid phone/FTTN network, requiring customers to arrange and pay for this themselves.
        • These cannot be DIY/self-install as they are on the Telco-side of the line.
      • Telstra standard charges for new phone or Bigpond "Ultimate"/"Velocity services requiring a technician visit are $300-$336.
  • The LNP Plan is silent on a critical point, nor has Turnbull ever answered the question:
    • Can a VDSL customer opt for an upgrade to pure-digitial service, with identical equipment to the other three Customer Access Networks?
      • All the other connection types include the installation, at no fee to the customer for standard installs, of an NTD (Network Termination Device) with associated Power Supply Unit (PSU), optionally with batteries. NTD's come with 2 Phone (UNI-V, or User Network Interface-Voice) and 4 Data ports (UNI-D).
        • These devices are the Telco interface and cannot be self-installed, a significant (>$300) professional fee will be incurred by someone.
        • To buy an NTD and PSU from a Telco or retail store will cost over $500.
        • The whole package would cost $750-$1,000 if provided by Telco's today.
      • Each of the ports can be provisioned independently and served by/billed a separate ISP/RSP if desired. This is very significant for Customers, Service Providers and Network owners, such as the Education Departments and Universities.
      • Is the LNP Plan for two classes of Customers:
        • No Fee Connections for Fibre (GPON), Fixed Wireless and Satellite versus 
        • Fee-for-Connection VDSL?
    • If a customer has already paid significant out-of-pocket expenses for a VDSL2 modem and Splitter, will they get a rebate for them?
  • Turnbull has stated, very clearly and on multiple occasions, that the copper network will be upgraded to full Fibre. This is not in the LNP Policy and its implementation problematic.
    • The timeframe suggested, usually in the context of Net Present Value (NPV) is 10 or 15 years.
      • Turnbull implies that with more advanced technology, this will be both much cheaper and more effective/efficient use of money. (Funds can be invested elsewhere until needed.)
        • The cost of electronics in the $1200-$1400 per service cost of GPON is only $300-$400. Most of the cost is on equipment, labour, testing and cable, none of which will decrease at all, or by much, in real terms.
        • We know from 30 years of PC's, that semiconductor costs don't drop much over time as new, higher-performance chips are released, soaking up Moore's Law savings.
        • At best, waiting 10-15 years will save $200, ~15%, but even that saving won't be realised. Real wages & cabling costs have risen much faster than this over the last 15 years. Waiting will increase, not decrease, the cost of FTTP.
  • The LNP has been silent on who would build this new replacement network for their VDSL/FTTN and how it would be funded, though it is inherent in their Policy, justifications and rhetoric.
    • The LNP have consistently stated they are ideologically opposed to Publicly funded Telecomms infrastructure works, especially the NBN.
    • The LNP have stated as part of their current NBN policy that they will allow and encourage competition for Telecomms infrastructure.
    • We know from both the 1995/6 HFC Cable TV roll-out and the post 2003 roll-out of ADSL2, that the large Telcos will aggressively compete leading to poor services, low coverage, high prices and much worse consumer outcomes.
    • The Fibre replacement for the LNP VDSL/FTTN network, based on the last 25 years of Telecomms in Australia, will be slow, expensive and fraught for consumers.
  • The issues of customer upgrades to FTTP via "co-funding" and "optional" full cost-recovery have been canvassed by the ALP. The LNP and Turnbull have not been forth coming in clarifying the actual costs or providing any certainty to Service Providers or Consumers.
    • Discussing these costs based on projections from the British Telecom broadband upgrade is speculation in the absence of any LNP statement.
The Government appears to have been working for the LNP: they've not challenged any of obviously manufactured estimates and fake figures, but have failed to release real data & estimates. The best response by the Minister for Communications has been to claim "it's all lies".

The Coalition Plan is to cover 9M services via a Fibre-to-the-Node (FTTN) network using VDSL2 with up to 2.8M services accessible via Fibre-to-the-Premises (FTTP) using GPON (Gigabit Passive Optical Networking). The rest of the NBN Co plan, Fixed Wireless and Satellite Customer Access Networks, CAN's, will remain the same. There may be more Fixed Wireless customers.

Whilst the Turnbull plan is very specific on the Capital Expenditure required (CapEx), $20.5 billion vs the $37.4 billion in the NBN Co 2012 Corporate Plan, it doesn't give any detail on how it's going to create those savings.

The only suggestion of its FTTN cost-model is "$900 per service" as an example for calculation.

The $17 billion can only be saved on the 9M VDSL services: an average of $1890 per service less.

Which is sheer fantasy, as the deployment cost of FTTP services averages only $1350 each, with another $850 per service going to Telstra as compensation for disconnected Phone Services, for an average cost of $2200 per FTTP service.

To reduce costs by the amount claimed, VDSL services must cost less than $310 each, including construction and Telstra contractual payments.

Either Turnbull isn't going to pay Telstra at all, a major policy shift and an outstanding commercial negotiation, or has found a way to shift all the CapEx costs to OpEx (Operational Expenditure) via leasing/rental agreements.

Only that can't be... Not only do the Coalition claim a $17 billion CapEx savings, they claim a 20-25% reduction in OpEx.

Each of these claims requires a complete suspension of disbelief, but together they are impossible.

Despite this catastrophic logic error, Turnbull has never addressed this question nor has his credibility with the Media seemed to have suffered for it.

NBN: The Coalition Policy Dud

As the first cab off the policy-rank, we, as voters, might have expected a cracker, a well researched and unimpeachable policy, from the Coalition and their tech supremo & one-time leader, Malcolm Turnbull.

What we got has more holes than a mosquito net, but more importantly, since its release the LNP (Liberal/National Parties) has not addressed serious defects and omissions, not updated their projections when better data has been released by NBN Co to its standing oversight committee, not released critical cost breakdowns nor answered fundamental questions, despite many requests to clarify. Turnbull did release a 22-point "FAQ": content-free and irrelevant is how it appeared to me.

Even if people have turned-off the NBN argument, it's not ever been a debate, or made up their minds on which version what they want, if at all, the behaviour of the LNP leaders should give serious cause for concern and reflection:
There is a pattern of intentional behaviour breathtaking in scope and degree that leaves well behind the Howard era's systemic and systematic redefinition & reversal of political speech and undertakings. This isn't a red-flag warning, it's the loudest of alarm bells.

This isn't merely simplistic "Policy by sound-bite", but a cynical, calculated misinformation campaign: nothing will be delivered as expected.
Sources: Urban Dictionary "non-core promise", Don Aitken on Lying vs Broken Political Promises, a tweet collection on the Howard Govt and its source: a 2006 speech by Sen. Faulkner and a 2010 election piece in The Punch.

Thursday, 6 June 2013

NBN: Aussie FoxNews agrees with Turnbull: "Connection to the NBN is _not_ Free"

What appears to be the on-line Australian version of FoxNews giving a unique version of "Fair and Balanced Reporting", Politifact, has given Turnbull a Thumbs Up for another of his outrageous misdirects, "Connection to the NBN is not free", made in what appears to be a complaint to the ACCC and first reported in Murdoch's "The Australian".  I don't agree with the Politifact conclusion:

Is it fair for Turnbull to conflate the cost of infrastructure installation and the cost of service charges, which are a long­established and accepted cost of internet access? They’re two very different aspects of the NBN. But as the ACCC points out in Advertising and Selling: "businesses may get into trouble with free offers if they do not reveal the complete truth." [emphasis added]

Turnbull says "connection to the NBN is not free". Regardless of Labor's investment, on being told they need to pay an RSP to get online many consumers would probably agree.

We rate Turnbull's statement Mostly True.
Others have discussed the overt-bias, funding & influencers behind the self-styled "Truth-O-Meter" of Politifact, so I shan't. It seems, like FoxNews, they use the same Recursive Fury and Self-Sealing Reasoning that Prof Stephan Lewandowsky of UWA has researched with Climate Change Deniers, self-styled 'sceptics'.

Lewandowsky & team provide strong research evidence for this, with an interesting wrinkle:
While you see/hear what you want to see/hear, the more educated or qualified your are, the more certain you are in in your views and stronger confirmation bias.
The Politifact article brings up the Trade Practices Act/Australian Consumer Law (TPA/ACL) issue of making "False or Misleading" statements. I raised this some time ago, and again, in connection with the Turnbull/Coalition plan. Whilst political promises are specifically exempt, Turnbull has been making very specific, even testable, claims about Commercial matters:
Will ASIC or the ACCC regard those promises are mere political puffery or commercial statements, misleading if false?
 I think Mr Turnbull in unwise in starting something that can get out of control. If the ACCC takes the view that the ALP statements are commercial, then they must also take the same view of Coalition statements. The differentiator may be what's included in political pamphlets and what's written/said in public media in support of a political policy with an implied commercial service.

The Coalition has failed to mention to potential customers significant and unavoidable out-of-pocket expenses they will incur with VDSL/FTTN and not with the other three, current, NBN Co transmission networks (Fibre, Fixed Wireless, Satellite).

Mr Turnbull has also very deliberately misled the public is stating the cost of VDSL modems will be about $50. There don't appear any stores in Australia selling VDSL modems even for $100, double that outrageous claim. Adam Internet supply VDSL Modems and CISCO ATA's for around $300. [Zytel site].

Meanwhile, in the UK - home of the VDSL/FTTN rollout Turnbull loudly spruiks, Amazon, generally known for competitive pricing, has resellers offering VDSL and cable/fibre router/modems nowhere near the A$50/UK£31 Turnbull claimed. The cheaper products are for cable/fibre, or ADSL, not for VDSL:
Lastly, what of the "NBN connection is not free" claim, how would an Ordinary Reasonable Reader (ORR) take the statements of the ALP and NBN Co? On face value, "NBN Co does not charge for standard connections, but does charge for non-standard" says charges, like phone connections, could apply but aren't normally levied. Whilst "NBN" conflates "NBN Co" and the many offerings provided by the multitude of RSP's using NBN Co's wholesale service over their three transmission networks.

Is there "an NBN", meaning just one National Broadband Network as implied by Turnbull?
Is there a standard price list for new "NBN" services offered by RSP's?
Do the RSP charges cover any physical devices or infrastructure?

No, No & No... There is a company with a similar name who offers wholesale, not retail, Customer Access Network (CAN.) services that are widely resold but are not the only CAN provider. The company doesn't have one network, it has three Customer Access Networks and a rather complex Transit network, with Interconnects for its customers and provides access to multiple, independent backhaul providers to RSP's who have to arrange their own hosting facilities, peering arrangements and access to "Tier 1" and international link providers.

There is no one NBN that anyone can point to. There are many wholesale services provided by many companies, including NBN Co for just one service, to a multitude of Retail Service Providers whom the public deal with.

The best we can say is that RSP's offer different services branded "NBN", mostly utilising Customer Access Networks provided by NBN Co. RSP's charge setup, not connection, fees for new services, even for customers of existing services. This is well known and understood by the public in the context of Broadband and other Telecommunications services, such as Mobile phones.

RSP's are at liberty to use CAN. wholesalers other than NBN Co and still brand their service "NBN". If wholesalers follow the NBN Co interconnect standards adopted under the ACCC SAU, Special Access Undertaking, RSP's can access customer premise equipment uniformly.

Bits are agnostic as to how they are carried: RSP's can lash together services from many wholesale providers and include their own product conditions and differentiators. Broadband Retail customers are acutely aware of the differences between RSP's and their packages and actively shop for retailers and deals that match their needs and values/priorities.

If there isn't "an NBN", then any assertion about "Connection to the NBN" is wrong before it starts.

For me, the question turns on what the general public perceive the NBN to be and if there's a difference between the usual wholesaler (NBN Co) and the Retail Service Providers (RSPs) that lash together many disparate services and components to provide a usable retail broadband service.

From the amount of press coverage and discussion since 2008/9, the public is very aware that NBN Co, the company, is involved in a major Telecommunications roll-out, that the major, but not sole, focus is on "Fibre to the Home", that Telstra is somehow involved in the roll-out and, most critically, they cannot buy anything directly from NBN Co: that they will continue to buy broadband services from ISP's, though they might be referred to as "RSP"s in documents. The detail and nuances of Network services and Regulatory environment are not nearly so well known.

Will people be charged when NBN Co come down the street and attach a Fibre  Premises Connection Device to their house? No.

When the customer orders a service from an RSP using NBN Co services will they be charged from installing the Fibre Wall Outlet, Network Termination Device and Power Supply Unit? No.

What is the "non-Free" charge that Turnbull is making much ado about?
It's a variable administrative charge by RSP's covering their service activation and billing costs.
[upd]. At least one ISP/RSP, iiNet, charges no setup fee, and I'm told others charge existing customers nothing to swap to an NBN Co plan.

This doesn't kill the "not Free" argument, but knocks a big hole in it. It changes the proposition to:
 "When connecting to the NBN, some customers may choose to pay their Retailer an initial fee."

Ordinary Reasonable Users of Telecommunications services understand these administrative charges.
The median age of the Australian population is around 45: almost all adults (voters) have had to pay for a telephone connection, landline or mobile. They understand that different charges can apply based on the amount of physical work to be done. That, at a minimum, an administrative charge applies.

There has also been enormous coverage, especially by the Coalition, about the cost of the NBN Co roll-out: people know that it's a bunch of money, many know that connections will average $3-4,000 each. Any administrative charge by an RSP of ~$100, I assert, will be regarded by ordinary, reasonable consumers as "being free", that is, they weren't charged for the equipment and lines installed for them.

Australian consumers of computing and Internet services are very used to "special" language in product claims. Microsoft for many years claimed versions of its Windows Operating System were shipped with "Zero" bugs. This was never challenged by the ACCC despite the overwhelming security and operation bugs in "Windows", acknowledged now in the monthly "patch Tuesday" updates released.
[Microsoft defined "Zero" as a commercially-tolerable level of defects.]

It's very well established in Trade Practice law that words such as "Free" and "Zero" can have non-literal meanings in specific consumer contexts. That consumers understand the normal caveats and conditions implied in advertising claims for specific markets and the intended meaning.

Will ordinary consumers feel misled or deceived when charged an administrative setup fee by an RSP for an "NBN" plan on Fibre, Wireless or Satellite? No.

Alternatively, how will those same consumers (75% of fixed lines will be VDSL under the Coaltion's plan) feel when they contact the same RSP and ask for an identical service, pure-digital with an in-premise NTD, on a Coalition VDSL service - to be greeted with a bill of probably $1,000 for equipment supply and installation? Very irate is my conjecture.

Note, this is not about the potential $5,000 charge for consumers optionally upgrading to full-fibre connection with NBN Co. The Coalition have disclosed there will be an upgrade fee, but not yet the scale of fees.

This is the game of misdirection and obfuscation that the Coalition are playing, being deliberately misleading themselves by not disclosing material concerns, whilst pillorying the Government over an innocuous and non-misleading statement.

Monday, 3 June 2013

NBN: Policy Ignorance by The Nationals Leader

Warren Truss, MP, appeared on ABC's Insiders yesterday and displayed an amazing ignorance of the Coalition NBN Policy and its implications. He consistently repeated their Magic Mantra "ours is Cheaper, more Affordable and Sooner" saying:
The advantage that we have in our system is that we will take our fibre to the node and as a result we will be able to deliver optic fibre cable connections to a much larger section of the community. Including to many regional areas
Only for people in the Country, the Nationals constituency, this is completely wrong... Fibre runs 40km-60km, not 300m-800m of Copper in an FTTN.

Nodes will be at least 800m from premises, which in the least populated/lowest density areas, means either very low speeds or expensive fixed wireless.
Unless he's assuming that Turnbull's Magic Plan will deem country households are "Cost Benefit" justified to get Full Fibre, presumably only if they voted for the Coalition.

Barrie Cassidy pushed Truss on the Coalition promise of 25Mbps connectivity to "every country household", asking:
How much will it cost householders to have access to that? [25Mbps broadband]
After repeatedly not getting an answer, he put to Truss that it would cost $6,000, which Truss rejected.

The Coalition won't acknowledge that their VDSL/FTTN entails significant ($500-$1500) and unavoidable out-of-pocket expenses for householders to a) get VDSL working at all and b) to have a real NBN Co NTD (Network Termination Device) installed.

Cassidy was presumably referring to an estimate of the "optional upgrade to full fibre" cost of the Coalition to overcome VDSL/FTTN speed and reliability constraints.

When you read what Telecommunications experts, not partisan Politicians pushing an agenda, write about the costs of deploying Broadband in Australia, and in Rural areas particularly, the answer is that the extended range

Prof Rod Tucker has multiple papers and presentations on the realities of Broadband, and strangely, Wireless and DSL/FTTN aren't the best or cheapest technologies.


A highly credible study by Ellershaw et al of Rural Broadband deployment costs for Wireless (WiMAX), FTTN (VDSL) and Full Fibre (PON) was published in the Telecommunications Journal of Australia in 2009, concluded that at 20Mbps, an FTTN was cheaper while at 50Mbps Full Fibre was always cheaper. The study chose 4 regional areas with different housing densities and 1 high density Metropolitan area. These telecomms experts used 300m, not 800m of the Coalition, to provide 50Mbps using VDSL. They also costs Nodes very cheaply at US$10,000 install + US$20/house in town and US$5,000 out of town. Current node cost estimates are $30,000 for upto 200 ports + $50/port for VDSL2.
Figure 13 Total cost of 50 Mbit/s broadband access
Passive optical network deployment costs are lowest for 50 Mbit/s access rate in all rural areas.

As expected, the total cost for both PON and FTTN DSL decreases as the population density increases. A comparison of the PON costs for the urban area of Burwood showed good agreement with the Verizon estimate of US$1350 per household by the year 2010 (Verizon 2006). [emphasis added]

Because a GPON can provide a dedicated 70 Mbit/s capacity to 32 households, the deployment costs of 20 Mbit/s and 50 Mbit/s are the same for PON. However, the FTTN DSL solution must use VDSL technology which is more expensive than ADSL. Since the range of VDSL is limited to 300 metres, the number of nodes required increases dramatically. Outside the rural towns, the households are often more than 300 metres apart and a VDSL node can serve only one household.

At this bit rate, WiMAX also becomes very expensive since only 16 households can be served by each base station.


Sunday, 2 June 2013

NBN: Hawke v Malby, Turjillo v Howard, Quigley v Turnbull, Ziggy v nobody

What makes Ziggy such a "good employee" that Turnbull wants him back to run a Government Telco?

The opposite of what got George Maltby fired by Bob Hawke at the end of 1988, Trujillo to spend his 5 years at loggerheads with Howard et al and Michael Quigley to be on Turnbull's "First to go" list:
To draw the ire of their Masters, all these heads of Australian Telcos exercised their discretion in doing their jobs and stood by their judgements.

What most voters may not appreciate is the extent that Government Business Enterprises (GBE) are directed by the Government and Minister. Sure, there is a Board, but they, and the manager, just follow orders. To do otherwise is cause for banishment - think of the number of ABC directors that have been "moved on".

In uncontroversial portfolios where profits are small, e.g. IP Australia, the Boards and CEO just get on with it. Telcos have always been highly political and seen by politicians as limitless pots of money. It's a great way to raise revenue without raising a tax or having to pass pesky legislation.

Government, not the GBE, decide the dividends, leading to Trujillo's 2005 criticism that Telstra had been dipping into Capital Reserves to pay its dividends. This is exceedingly poor business practice that in a private business leads to eventual insolvency.

Maltby had similarly complained in the 1980's that O.T.C., "the most profitable public utility", was being required to pay too much money back as dividends, forcing it to borrow more than was necessary. Governments for years had milked these cash cows. As O.T.C. steadily increased both its revenues and profits, while decreasing call costs every year, through the decades of the boom fed by Satellite & Cable improvements, successive Governments availed themselves of this newly found piggybank.

In August 2005, just five weeks on the job, Trujillo lobbied Howard and his senior ministers for significant changes. Telstra was in serious trouble, its income dependent on a failing business, Telephony. In 19 slides, Trujillo devoted to illustrating his point: Telstra was in serious difficulties, had been for some time and things would only worsen. Was he banging the table? I would've...

Trujillo didn't leave it there, he went with a plan to fix the problems: create an NBN with a FTTN network.
He thought they could be finished by the end of his contract, 5 years.
To meet Telstra and normal "Traditional" Telco financial targets, Telstra needed regulatory changes to allow them to charge enough and control the market in ways they wished. Ostensibly, this withholding of this regulatory change was what created the rift between Trujillo and his team and the Howard government.

Could it though?

Telstra was in deep trouble that had been worsening for at least  5 years. We know from Thodey, the current CEO in a 2009 briefing, that in 1995 Telstra had seen the writing on the wall and formulated a commercial strategy to replace the copper PSTN (Public Switched Telephone Network) by 2010. This telephony only Fibre to the Node network would've been cheaply upgradeable to ADSL broadband, just as Telstra have done with their "Top Hat" programme.

Telstra have executed part of that strategy with around 8,800 RIMs & CMUX's deployed, serving an estimated 1-2M lines.
Why weren't successive Telstra CEO's allowed to execute this strategy fully, given it could've been executed within its usual financial envelope?

A first guess is two-fold: a) Governments took excessive dividends putting the investment out of reach and b) successive Governments directed the network not be replaced.

When Howard sold Telstra in late 2005 for $3.60 a share, well below the $7.50/share of T1 and under the $4.50 anticipated in the Budget. Do you think Telstra would've been worth more or less with a well advanced FTTN? I think more, at least the $7.50 of T2, probably 20-30% more.

Could the regulation of that FTTN network been handled simply?
It was obvious, possible and trivial right up until the T3 sale: split Telstra in two corporations, one wholesale only and the other retail. Perhaps with a third corporation with all the mobile phone business.

The Howard government didn't just stuff things up, they created a nightmare that is still far from resolved and so the sake of maybe $2 billion invested in 1999, robbed themselves of $20 billion in increased value.

This is where Trujillo came in.
The Howard Government went out and deliberately recruited the best Telco executive team they could find in the world, then argued with them when told there were problems and how to fix them!

Could Telstra CEO's since 1995 not have been telling the Howard Government, in power in early 1996, that the organisation was in deep trouble, needed to fully execute the "replace the PSTN" strategy, that they could do it cheaply and it would add significant value to Corporation when it was privatised?

I don't believe any competent Telco CEO could have failed to notice the structural shifts and business challenges and know the solutions available. I also think that Ziggy Switkowski, a PhD in nuclear physics, was smarter and more perceptive than most managers and didn't miss this change.

Ziggy also planned all three tranches, T1, T2, T3, of the Telstra sale. I can't believe the Telstra economists and actuaries weren't also commissioned to run forecasts of the impact on the sale price of investing, or not, in a PSTN replacement. If that's the case, the timing of the T3 sale in 2005, well after the 1999 T2 tranche, may have been chosen by Ziggy. The share price fell steadily from the highs of the "Dot Boom", but still stabilise. Were the forecast models out by just 12-18 months?

Did the internal Telstra models show a positive Net Present Value as you'd expect by removing a huge maintenance overhead and creating multiple new lines of business with good Gross Margins?
Did these models get presented to the Board and the Government as you'd expect?
And did they go onto the Howard Government?

All we know for sure is two things:
  • The "replace the PSTN" strategy was never fully executed, creating a predictable financial crisis by mid-2005 when Trujillo arrived and 
  • Ziggy said nothing publicly about this aspect of the business at the time, nor since.
In many eyes, this makes Ziggy "The Perfect Employee", enough it seems to get him pre-hired by Turnbull as CEO of NBN Co.

What do Sol Trujillo and Mike Quigley have in common?
Before coming to Australia, they both  ran large, global companies involved in Telecommunications, known for aggressive competition between companies and for "very robust" internal politics and "no bull" cultures.

It seems that being very competent, straight-talking executives who will robustly defend their views is anathema to Australian Governments. We know from Maltby v Hawke this is not solely a Coalition problem.

Quigley has another feather in his cap as far as I'm concerned: he didn't need the job, as demonstrated by donating his first years' salary to charity.

There are two things in there - he took the role on because he was prevailed upon and he has a strong moral centre which he won't compromise. Donating a $2 million salary to your favourite charity is the mark of a very secure and committed individual who won't compromise their principles and beliefs.

I guess that's another reason Quigley and Turnbull don't get on.

Friday, 31 May 2013

NBN: Conroy responsible for maintaining the politicisation of the debate

The choice of Customer Access Network (CAN) technology and the means of its update are Technical, Commercial and Economic considerations. These are dry, rational discussions, based not on opinion, but hard facts and numbers. Bits are completely agnostic: they work no matter what network or connection they arrive over.

Instead, we have the mess that's the politicised NBN Argument of Ideologies. In no way is it a debate, but that could be changed somewhat by a single action: an independent review of the NBN Plans of both sides.

This politicisation only brings heat, not light, and doesn't help the electorate decide on facts, but forces them to take sides. This can only lead to a poor result, failed expectations and years of recriminations.

This craziness didn't happen for much more important issues: the switch-off of analogue TV, unleaded Petrol and change to Metric System. What's so special about replacing our communications infrastructure?

The Government, by not damping down this sideshow, is allowing a bad situation to become worse.
Very little is needed to take the wind from the Coalition arguments and deny the rabid trolls their oxygen:
  • hard data, credible numbers
  • independent, expert analysis of the strengths, weaknesses of designs and sound financial modelling of Project Risks, both upside and downside, of both major party plans.
    • And Apples-and-Apples costings, run for more than the life of the transitional VDSL/FTTN network.
Turnbull's policy modelling included stress-testing of the NBN Co plan. While I think that the figures chosen are unsupportable and selected only because they could produce a bad result, the approach & attempt is one of the best policy reactions by any Opposition I've seen in 40 years.

That's an important point: Turnbull must be lauded for adopting a thoughtful and credible response to the Governments' roll-out. Project stress-testing is exactly the right approach. Outrageously fiddling the numbers demeans and devalues the analysis in my eyes.

Senator Conroy as the Minister for Communications has the money and commercial-in-confidence data available to create an authoritative analysis of both policies. Useful models don't have to be finely detailed, only "good-enough", with, say, a 5-10% error. An informed, capable team should be able to prepare a useful report in 3-4 weeks, provided they are given access to the required data and esimates.

This is what I'd like to see done by independent experts given access to the data & estimates needed:
  • Turnbull's four stress-tests rerun with believable data and a stated probability level. (1 in 10 or 1 in 100)
  • Turnbull's Worst Case Scenario rerun with new data and its likelihood stated (1 in 100,000?).
  • The same, or appropriate, stress-tests run for a VDSL/FTTN network.
    • Including a detailed CapEx and OpEx breakdown until 2040, matching the NBN Co Corporate Plan.
    • The impact on Revenue of "single access rate charge", cherry-picking and "ACCC mandated highest wholesale charge".
  • Full Fibre and 70% FTTN models both run for the current NBN Co plan, with Download Volume growths of:
    • Current 30% pa
    • ABS's 50% pa
    • CISCO VNI 66% CAGR
    • And the current 44% 100Mbps with average 47GB/mth, not 30Gb/mth as the baseline.
    • Update for 1Gbps service. Unsure adoption rate to use.
    • Consumer Out-of-Pocket expenses needed for first "NBN Connection" point.
      • Line testing and remediation costs, if any, needed for VDSL.
      • VDSL+Telephony needs a Central Splitter
      • Standard 2*UNI-V, 4*UNI-D NTDs (Network Termination Devices) are not mentioned in the Coalition Plan, while they maintain "In-premises costs are not included". This needs clarification.
  • Under the 70% FTTN model, the break-even period built into the Coalition model and the conditions that might lead to the Copper CAN to be shutdown before this break-even period.
Experts in this field will come up with other, better, tests and modelling.

Also, Telstra need to issue an ASX statement clarifying Mr Turnbull's insistence "They will" in answer to "Why do you think Telstra agree to your proposal". If there have been back-room deals done, which I believe not to be the case, Telstra needs to inform the market.

Mr Conroy won't listen to me, I don't have recognised credentials in the area. Perhaps he'd respond to a petition or open-letter from a large group of Australian Telecomms Experts, including, say, Prof Reg Coutts from the original NBN Expert Panel.

Anyone out there able to put this to any Telecomms Experts?