Turnbull to Kohler: Node, node, node

Alan Kohler's latest counterblast on the NBN criticises me for going to the election "with a plan that won't be very popular, based on saving money.” He goes on to say: "Worse, the money saved – Turnbull estimates $20 billion – can't be spent elsewhere or used to bring down taxes, because it is capital expenditure, not operating expenditure” (Sorry Malcolm, it's still crazy, August 21).

Alan is of course entitled to give me political advice (why not, everyone else does) but it is surprising that a financial writer would be giving such cynical political advice and paying little or no regard to the responsibility all of us have for spending taxpayers' money prudently. He seems to be advising me to be like the Labor Party – spend public money with only one objective in mind – political advantage.

Alan initially massively overstated (by twentyfold) the likely extent of connections by the time of the next election, but I gather his argument is now that there will be many premises in areas covered by contracts to build fibre to the premises by the time of the next election and that it would be expensive or wasteful or both to cancel those contracts.

I am working on the assumption that we will honour contracts which are in place. This does not preclude renegotiating or varying contracts where that makes commercial sense of course.

So, yes, it is very likely that in urban Australia under a Coalition-managed NBN there will be areas with FTTP and areas with fibre to the node. Over time the proportion of FTTP will naturally grow as FTTP will be deployed in greenfields developments. New apartment buildings will most likely have the fibre terminating in the basement at a DSLAM connecting into the copper LAN in that building as is the practice in most countries, including Korea.

And as I noted yesterday, there is nothing remarkable about having a mix of technologies. Most of us use the internet over different channels every day – 3G or 4G wireless, wifi, ADSL over copper, HFC, fibre to the premises (Alan Kohler's NBN fantasy, August 20).

Alan argues that the people in a FTTN area will be bitter and aggrieved because they do not have FTTP. I disagree. While politicians, journalists and geeks write endlessly (and tediously) about this stuff, the vast majority of Australians are only concerned as to whether their connection enables them to use and enjoy the services and applications they want.

My contention, based on experience here and abroad, is that FTTN has the capacity of delivering very fast broadband which will meet those customer needs. Obviously business centres, schools, hospitals, universities and so on should have FTTP (many do already).

Alan's other point is that it costs more to maintain copper networks than optical fibre networks. There are a couple of points to bear in mind here.

First, remember that the copper on the E side of the network (i.e. from the exchange to the node) will be entirely replaced with fibre. Therefore the only copper left will be on the D side (from the node to the customer's premises).

Second, where there are chronic maintenance problems (such as very damp ground) it may make economic sense to run FTTP.

Third, while I accept that as a matter of principle optical networks should require less maintenance (work by Analysys Mason for the UK Government's Broadband Stakeholders Group in 2008 estimated FTTP opex was approximately 30 per cent less than DSL opex) in discussions with CTOs of large telcos over the last year or so I have heard a rather different story. Their experience is that the maintenance argument put in favour of fibre is exaggerated.

As one of these very experienced chief technology officers, who is deploying both FTTN and FTTP, said to me:

"Fault rates are not significantly lower with fibre than with copper in our experience. Fibre is very sensitive to the installation procedure, the fibre plant may be fine up to the E side but from there on, the D side, to the premises there are risks. A copper cable is inherently more robust and more forgiving of contractor errors and mishandling. With fibre plant you need to ensure the contractors pay a lot of attention on installation. It is easy to make expensive mistakes.”

Another CTO, from North America, told me that in his experience maintenance on outside fibre plant was about 50 per cent less than on outside copper plant and that they worked on an average maintenance cost for copper connected customers of $US3 per month per premise – so $US36 a year or about $US18 a year more than for fibre. He said that the additional capital costs of FTTP so dwarfed the opex savings (they don't even pay the interest) that the additional maintenance costs of FTTN were not a significant factor.

Now before FTTP enthusiasts explode with indignation, bear in mind I am simply reporting what I have been told by people expert in the field and running very large telecommunications networks.

It is also worth noting that in the latest NBN corporate plan there is $1.5 billion to $3.0 billion allowed for annual capital expenditures in the years after the network has been built. It is not clear how much of that is maintenance, but in the previous Corporate Plan most of this expense was ascribed to ‘replacement' capex (this figure is no longer broken out).

A final point on maintenance: the TUSMA contract with Telstra to maintain copper in the last 7 per cent of premises has an annual expense of $290 million, with approximately $230 million of this directly attributed to the cost of upkeep for this part of the copper network. Given most estimates of the national maintenance cost of the copper are around $750 million per year, Stephen Conroy's decision to sign this 20-year contract has locked a third of these costs in for a generation.

Returning to Alan's latest article, I am surprised that he takes seriously the rather Orwellian NBN Co statistic of premises where construction is "commenced or completed”. This is Orwellian because it tells you nothing about when the construction will be completed – and what does "commenced” mean? A shovel in the ground? A plan drawn up? The only metrics that are useful are premises connected and premises passed (and therefore capable of connection once a customers asks for a service).

His conclusion is that a change of policy to save tens of billions of dollar is "Lots of pain, little gain”. Apart from the taxpayer of course who gets to save tens of billions of dollars and internet users who get their upgrades a lot sooner and, because the capital investment is so much less, more affordably.

The Coalition's approach to the NBN Co is not political or ideological, but rational. We will complete the NBN, there need be no anxiety or uncertainty about that, and we will do so sooner, cheaper and more affordably for users.

Malcolm Turnbull is the Member for Wentworth and Shadow Minister for Communications and Broadband.