Another NBN rant

There are a couple of things that the NBN fanbois generally fail to note about Liberal Party Broadband Policy.
(and I don’t vote, so this is just a technical & economic discussion)

It is not about FTTN at all. FTTN is only a cheap and fast “filler” for areas of Australia that aren’t being done some other way.

  • Satellite remains.
  • Fixed wireless remains.
  • Greenfield FTTH remains.
  • FTTH On-demand (i.e. anyone who wants to have it) remains.
  • HFC (and we know that EuroDOCSIS can do 1Gb/s services, as needed) will be recovered and provide service for 2.6 million homes, immediately.

All that is changing is that the difficult brownfields services are being given a fast and cheap alternative to get some kind of service upgrade in the short term (say 5 to 8 years).

Fanbois bitch about the FTTN cabinets in the street. Yes they’re ugly. Essentially, they’re providing centralised networked power to drive the copper tail for all FTTN subscribers. Cabinets are ugly, but they’re very maintainable, and upgradable at low cost and with low impact for consumers. The current FTTH solution is putting an unmaintainable ugly lead acid battery IN MY LOUNGEROOM! (And, forcing me to pay for the power for their service too. I’m used to getting my POTS power for free.)

And, if (when) the FTTH battery is dead my lifeline services are affected. It is part of every NBNCo service agreement that it is not their liability, if your battery is dead, you can’t make a call, and therefore you die!

I’d rather have someone professional maintaining my (mother’s / great aunt’s / disabled friend’s) lifeline services battery and have it somewhere where it can be properly maintained to deliver a known grade of service, without interfering with my life.

Oh, and the famous diagram with 1Gb/s FTTH services on it, with everything else miles below. Nice picture. He could have added HFC services up there at 1Gb/s too, but he didn’t. But paying for these 1Gb/s services is another story.

The NBNCo has gazetted their 100Mb/s prices in a SAU with the ACCC, and their prices will be INCREASED annually by CPI (-1.5%) for 27 YEARS. That means in 27 years we’ll be paying much more than today (compound it up, I dare you) for the same service. They need to lock in this price to pay for their network. Don’t imagine that Moore’s law is going to apply here. 27 years ago was 1986, and then 9600 baud was looking pretty good. How will 100Mb/s look in 2040? Now, how much is then 1Gb/s going to cost, if 100Mb/s has a fixed price for the next 27 years? Good question. You got any unwanted children you can sell?

UPDATE: So just a day after this rant was written, NBNCo announced pricing for 1Gb/s services, conveniently only available after the next election in December 2013. Pricing is $150 wholesale. This is exactly twice the 100Mb/s price. An interesting price point, given this will cut the fan-out on the GPON network infrastructure from 32:1 down to 2:1. Also RSPs will need to upgrade their POI backhaul packages by a factor of 10x if they want to provide this service. Using the ACCC transmission pricing calculator, this looks like a recipe for a VERY expensive retail service.

The Labour NBN plan is unbuildable. There are not the fibre splicers in the country we need to achieve the required daily rate build rate. NBN Co contractors (to get their contracts in the first place) are paying lowest rates in the market. Anyone with any skills is working on the mines in WA, fly in fly out. Not camping in the truck, schlepping around the country digging trenches. This was apparent back in 2010. Blind Freddy could have predicted the situation the NBNCo is in now, with missed delivery targets.

In 15 years, Conroy might be remembered for the man who thought of NBN. But Turnbull will be remembered as the one who saved it, and delivered it.
(Actually, Conroy should be remembered as the man who killed NBN, when he axed its predecessor OPEL, on April 2, 2008, when he entered office).

UPDATE: So on 12th July 2013, Mike Quigley resigned or retired from his role as CEO of NBN Co. Surely his efforts will not go unrewarded. My bet is that in the second half of 2014 or early 2015 Mr Quigley will be appointed to the Board of Alcatel Lucent. After all, getting your former employer an uncontested 1,500 million Dollar contract from the Australian Government for obsolete GPON equipment must be worth something, right?

Our NBN; make it better, please.

I’ve written my personal thoughts on the difficulties in creating a successful outcome for the NBN in Australia previously  in Our NBN is an Engineering Fail.

And today I made some comments on the “jumping the shark” article on ZDNet, which got me soundly voted off this small Island. Jumping the shark.

But, before I pack my bags and depart, I thought that I should leave some positive words on what the NBN Co could do better, and remove my humble objections to the path they’ve taken.

1. Commit to a public “rising tide” deployment strategy.

The principal of providing broadband access to all of Australia is commendable, and should be pursued. But, the current deployment strategy is opaque and leaves us open to believe that it can be gamed for political grounds.

Secondly, it is clear that as a result of the existing deployment strategy, existing functional high performance broadband networks are destined to be switched off earlier than necessary.

The deployment strategy should pursue the path of providing broadband to those with no broadband first, before any customer with an existing DSL service is overbuilt. Once this is achieved, then those with a ADSL 1 or other slow or older broadband technology should be over built.

Radio and satellite deployment schedules should clearly be continued.

Obviously this strategy would reach those with ADSL2+ next to last, and those with HFC network access last of all. This is a good thing, as it would focus the initial effort and money on the areas where no network service currently exists.

This deployment strategy would allow the average national broadband speed to float quickly upwards on a “rising tide” of performance, as the lowest performance outliers are removed.

 2. Devolve all vendor contracts into contestable segments of no greater than $100 million.

There is no technology that can’t be devolved into contestable packages of around $100 million or less. The idea that the NBN Co can sign off exclusive contracts for $1+ billion dollars, without any competitive options, and without sufficient in-house resource for intense contractual performance scrutiny, is just lunacy.

Experience shows that smaller contracts ensure that deliverables can be more closely aligned with outcomes, and are much less likely to have significant cost over runs. Agile delivery of outcomes is required here.

3. Establish competitive tension in procurement.

The NBN Co should ensure that each and every technology package is procurable from at least 2 geographically diverse technology vendors, with demonstrably different supply chain mechanics. It takes just one geo-tectonic event (Icelandic dust clouds, or tsunami for example) to lay waste to an entire geography for months. Disruptions to a single vendor’s supply chain should not have a downstream effect on every supplier to the NBN and the resulting build schedule.

Vendors also sometimes take mis-steps in their development processes, leading to obsolete equipment being placed into the field. Over a 10 year build process, it is relatively difficult to keep watch that procured equipment remains fresh, when no competitive tension is in play.

NBN Co should ensure that two “equal” vendors are selected for each and every contestable product and service segment, and that no favoured vendor is driving the NBN Co architectural outcomes.

 4. Recognise NBN Co is a utility and national standards provider and rewrite their business plan based thereon.

The political expediency to have a “positive” business case has shackled the NBN Co to an unrealistic build schedule, and elevated long term wholesale costs to Retail Service Providers.

The NBN Co should be recognised as a utility, and our wholesale network provider of last resort, to provide a broadband service where no commercial service provider is prepared to invest for a commercial return. And in return, the NBN Co should be freed from the need to cross subsidise from low cost urban deployments to higher cost deployment regions.

NBN Co should be financed from general revenue, in the same way that other major infrastructure projects are financed.

NBN Co should tender for operating companies to build and operate geographic sectors of the Australian continent’s fixed wholesale network infrastructure (based on the “rising tide” principal above) to the standards set by NBN Co, to secure the maximum involvement from private enterprise.