Technology and Games: Content from Across the ABC

How does this site work?

Blog

Huge doubts cast over Coalition's 'cheaper' NBN alternative

Nick Ross ABC Technology and Games 16 Nov 2012
Light streams through fibre optic cables

Have we reached the tipping point - the point of no return? Credit: Rodolfo Clix, file photo (www.sxc.hu)

Comment

[Update: there appear to be some issues between what NBNco and the contractors (referenced in this article) count as being a 'connection'. That's in the process of being ironed out. Nonetheless, the general themes below all stand.]

[Update 2: Jim Madigan, NBN Account Manager at SPATIALinfo - a Victorian contractor for NBNco - confirmed to us that what the contractors said at the conference stands - albeit with a slight variation. To clarify: for the first half of next year the contractors will be passing 6000 homes per day. For the second part of the year they will be passing AND CONNECTING 6000 homes per day. This means that the rollout speed is EXCEEDING the aims of the NBNco corporate plan which aims to pass 6423 premises per day in 2015. This will doubtless be discussed further in other articles but for now, the primary points of this article stand.]

Of the many broadband forums in the world, NBN Realized, which ran in Sydney yesterday, may not be the most glamorous. It focuses on the actual building of the network and was attended by the people who are on the ground installing cables to the houses, over and under the streets along with the people who make the cables and networking infrastructure. While no one wanted to go on the record directly, there was consensus from key players in the room: building a Fibre to the Node (FTTN) infrastructure, in terms of raw construction costs, as promoted by the Coalition, is now unlikely to be a "cheaper" option than the current Fibre to the Premises plans.

There are already many arguments to counter claims that FTTN is cheaper: the higher running costs, loss of savings to existing infrastructures, operating models, boosts to business and GDP etc all mean that FTTP is substantially cheaper when the major costs and benefits are factored in. But none of these arguments could counter the notion that the basic, cost-on-the-ground, building costs of a FTTN-based, basic infrastructure were cheaper... until now.

In a nutshell

In essence, NBN contractors such as Silcar/Thiess, SPATIALinfo and Service Stream are now so efficient at rolling out fibre down streets - from the exchange to people's houses - that stopping to add tens of thousands of large-fridge sized node cabinets represents an expensive, time-consuming hindrance. The hindrance comes with a practical nightmare of powering the node and intense bureaucracy born from the requirement of dealing with power companies, associated regulations and the necessity of using power companies' own engineers to hook each one up. One contractor said, "I don't know how you'd power them."

The contractors were understandably unwilling to go "on the record" and say that a political policy was cheaper. They also pointed out that it could still be cheaper if the scope of an alternative rollout was reduced. Nonetheless, they did mention that the state of the copper would be an issue. Other infrastructure leaders subsequently concurred and pointed out that maintenance to Australia's copper networks (both PSTN and HFC) has practically collapsed over the past three years because there has been little reason for commercial companies to spend money on maintaining networks that are scheduled to be ripped out of the ground. There's now a very real question of whether it would be feasible to rely on any infrastructure, which relies upon existing copper, even if we wanted to.

There's more

The contractors foresee no problem hitting 6000 connections per day from early next year which would mean NBNco's plans are running on schedule. Stephen Ellich, Director of Service Stream, explained how the rollout process required passing through various phases, "The first phase is design. When everybody started everybody was probably thinking, 'Well we'll never design that many jobs at once!' Now, you know, we're simultaneously designing jobs in the order of...
30,000 to 40,000 premises per month."

Dan Birmingham from Silcar chipped in, "We're hitting 50,000 a month at the moment."

Ellich went on, "So you add those numbers together and the puzzle starts to come together.

"We're ramping up each phase of this project to build it. It's not like it's a 'day one' requirement - we see it truly as a journey. And it's that investment in the resources, training, and the infrastructure which includes our own plant out in the field, our own splicing equipment, our own CAD stations, spatial access and training, that gives us the ability to build an industry that is capable of building 6000 at every phase."

Tony Cotter, MD and CTO for SPATIALinfo pointed out that the rollout was not unprecedented and that when Telstra and Optus rolled out their HFC networks, they were hitting 5000 per day while in the US the 'Verizon - FiOS' roll-out was hitting 5000 to 6000 connections per day

"It's not unachievable" he said.

Ellich stressed on several occasions that he didn't know if FTTN would be cheaper but he did say the following, "I don't think it's simple changing rom Fibre to the Premises to Fibre to the Node. For example, you have no electronic cabinetisation that exists out there as a start. So, if you went to the node, and where those nodes might have to be, you've still got to change from one media to another which requires some electronics out there. It would require power to those electronics, presumably.

"I just don't think it is as simple as people think; that you just, all of a sudden, stop providing fibre in the last three-or-four hundred metres or maybe even the last kilometre from where it might be a pillar, for example. There's... a whole bunch of work that will happen that's not currently required with this rollout."

Birmingham added, "The big question is just that if you go with fibre just to the node, does that node become active [it does - ed]. If it becomes active then you've got to connect power to it and that's a significant cost."

Cotter said, "You're also relying on the existing copper network from the node to the premise. And I believe Telstra stopped rolling that three years ago."

I asked, "Can you tell us what state the copper is in?"

"I wouldn't tell you on the record, no."

Ellich backed him up, "It's probably not up to us to comment on." This is disappointing considering that these are the guys ripping it out of the ground. Nonetheless, this issue has been around for some time and directly affects the viability of any future infrastructure right down to its fitness for purpose. It needs to be answered at some stage should the NBN be halted.

All this means that, by the time of the next election, the fibre roll out could be a hugely-efficient process that would have been in full swing for the best part of a year.

Nodes

Cabinets like these would need to line many of Australia's streets if Fibre to the Node was implemented.

There are other issues related to connecting power to the nodes. At present, fibre doesn't require extra power to connect from the exchange to people's houses - signals can travel many kilometres without the signal degrading or needing amplification. However, nodes need to be powered and installed with multiple backup batteries. One contractor told us that 65,000 nodes may be needed and that powering them would represent an "environmental disaster". Indeed, Rod Tucker, at the University of Melbourne, has said that this would require the construction of two-to-three new power stations. These would come with a substantial cost of construction and also running costs.

Old copper

What also became clear from talking to the expert conference delegates was that the owners of Australia's copper networks, principally Telstra and Optus, have all but given up on maintaining their networks in the "three years since NBN came knocking". As one cabling specialist told us, maintenance has been dropping off significantly since the 90s and, while such practices aren't ubiquitous, overworked Telstra workers have focussed their maintenance operation on fixing things when they break rather than preventative measures. As such, when problems have been fixed - the symptom and not the cause has often been dealt with exclusively. An example given was that a line would be rotting due to 'moisture getting in' and so the line would be replaced but not the issue affecting the moisture. In the past three years all but essential maintenance to Telstra's copper PSTN network and Telstra and Optus' HFC networks has collapsed.

Another major issue was the reminder that in order to physically build FTTN, one would also have to buy or lease the copper from Telstra at a capital purchase of between $4bn and $20bn depending on who you choose to believe (Alan Kohler and Malcolm Turnbull had an entertaining debate on this here:
The Coalition's NBN policy is madness
Alan Kohler's NBN fantasy
Sorry Malcolm, it's still crazy
Ultimately, things boil down to how nice you think Telstra would be when selling or leasing their unique asset to the government as part of an election promise.

More realities involved the maintenance of nodes. Water is the enemy of copper and corrosion decimates network performance - some people can't even make phone calls. With FTTP the electronics are housed safely in the exchange at one end and at a person's house at the other. With FTTN, the cabinets are once again at the mercy of the elements and will require capital expense to have maintenance vehicles driving round the country maintaining them - an enormous cost. It should be remembered that it already costs over $1bn every year just to maintain the copper network.

A major infrastructure provider also reinforced the problems of connecting fibre networks with copper. Several types of software are required to get the two networks to talk to each other and exchanges will have to manage both. This was described as a "nightmare".

As for who would manage and own the nodes is another question that needs to be answered. Telstra has previously stated that allowing other companies access to its nodes would be a "nightmare" in terms of integrity and management.

More obscure benefits were discussed by some contractors. Paul Walker of Silcar described the effect of the NBN in ways no one had yet thought of. In some streets in Melbourne two sides of the street see unkempt pavements lie opposite rejuvenated pavements where NBN contractors had been and left. In addition to this, telegraph poles used by NBN contractors have also been rejuvenated - they come with a fixed lifespan and need to get replaced at a cost of $20,000 each. NBN is responsible for helping with many such replacements and this alone could contribute "a billion dollars" towards existing infrastructure maintenance efficiencies.

Ultimately, however, the message from the conference was clear: despite nobody wanting to be on the record saying it: reversing the rollout or even stopping it is now considered by many of the people who know, to be unthinkable purely on a practical level.

Previous discussion has revolved around a tipping point whereby the NBN can't be halted and it was assumed that this would happen around the time of the next election, if at all. Now however, one major infrastructure player said, "We've reached the tipping point."

Even if we have to wait until 6000 premises are being connected per day to validate these claims, the fact that this is scheduled to happen early next year means we'll likely know sooner rather than later - and almost certainly before the next election.

Previous discussion of delays has been muddied by politics and focussed on mixing and matching early NBNco projections with local rollout issues. The discussion is arguably going to become moot very soon. If 6000 premises are being connected every day from early next year (meaning that NBNCo will hit its target), what point is there discussing the delays preceding it?

Blowouts

Another factor discussed by the delegates was the notion of cost blowouts. Much of the conference dealt with experience garnered from actually installing equipment in various premises and learning about what to do and what not to do. Installing boxes in houses where small children could access them, or people could trip over power cords that trailed across floors because the power socket was some distance away, had led to developments in installation best practices. Issues were discussed with regards to dealing with the placement of boxes in heritage-listed houses and how many people were choosing an area near the television for the main box to be installed in order to best make use of IPTV. Practices were improving in efficiency across the board and as a result the chance of massive cost blowouts were dwindling.

Mike Kaiser, NBNco's Head of Quality, added to this by showing how the Cost Per Premises of the Tasmania rollout had already seen improvements when moving into phase two; "As we improve our processes, as we drive cost and waste out of our processes." Further cost reductions had been made with the Mainland First Release.

Kaiser was adamant that by 2013 the cost per premises will be where it was stated to be in the NBNco corporate plan. (As a final addendum, Kaiser also said that NBNco fibre customers use an average of 65GB of data per month compared to 17GB for DSL users. This more-than-tripling of data consumption is worth more attention.)

Is it real?

It's worth stressing the point, that there hasn't been any serious audit work to confirm all of these claims and figures. However, as a whole, with similar information coming in from multiple sources and often being backed up by figures and process descriptions, it now represents the best information in this area that we have. Importantly, what is now being said makes sense. If indeed 6000 premises per day start getting connected early next year, then arguably we'll have our answer.

If that happens then it may be fair to say that moving to a Fibre to the Node-based infrastructure is as untenable as moving to a wireless based one.

Politics

There's much to argue about politically, and I'm going to step back here as it's not my domain.

But there's no escaping the fact that what's above goes directly against what Joe Hockey recently told the ABC:

"There is not one contractor in Australia that believes the Government is going to roll out its National Broadband Network for $32 billion. Expectations are as high as $60 billion, $70 billion or even $100 billion for the National Broadband Network."

It also goes against Malcolm Turnbull's recent statement to Leigh Sales that "The difference between fibre to the cabinet and fibre to the premises cost differential is about four to one."

Ultimately, we find ourselves again in the situation of asking the Coalition if we can see the details of how their policy works in order to analyse it and draw conclusions.

[an error occurred while processing this directive]

How Does this Site Work?

This site is where you will find ABC stories, interviews and videos on the subject of Technology and Games. As you browse through the site, the links you follow will take you to stories as they appeared in their original context, whether from ABC News, a TV program or a radio interview. Please enjoy.

ABC iView iPhone App

ABC Download This Show

ABC The Brief iPad App