Study: More fibre, less retail cost, for rural NBN users

By Liz Tay on May 6, 2010 5:02 PM
Filed under Telco/ISP

Implementation report recommends wider coverage.

A comparison of Government’s initial cost estimates and “detailed cost modelling” conducted by ...	The projected internal rate of return (IRR) for NBN Co is sufficient to cover the cost of ...	From a commercial standpoint, Government is the most natural owner of the risk versus return ...
See all pictures here »

More rural Australians could receive fibre National Broadband Network (NBN) access and at a metropolitan price point, if the Government heeds recommendations made in the NBN Implementation Study released today.

While the Government planned to deliver 100 Mbps fibre connections to only 90 percent of Australian households, authors of the $25 million report have called for fibre coverage to be extended to 93 percent.

Report authors at KPMG and McKinsey and Company expected it to cost $2.3 billion more to deliver the additional fibre connections instead of the 12 Mbps wireless connections that had been planned.

But although fibre cost more to deploy, it was likely to have a higher uptake -- and cost was not prohibitive, the study found, highlighting lower backhaul cost estimates than those initially anticipated.

Implementation study estimate of cost per premise to deliver a fibre connection - Source: KPMG-McKinsey

Initial Government estimates priced the NBN at $43 billion, allocating $28.5 billion for fibre to 90 percent of premises, $3.0 billion for the remaining 10 percent, $5.4 billion in backhaul costs, $1.5 billion in overhead costs and $4.6 billion for out-turning.

The implementation report budgeted $26.6 billion for 90 percent of premises, $5.3 billion for the remaining 10 percent, $3.3 billion for backhaul costs, $2.0 billion for overhead costs and $5.6 bilion for out-turning, which came to a total of $42.8 billion for the NBN.

Of the seven percent of premises that did not receive fibre connections, four should receive a commercially-tendered fixed-wireless service with a wholesale Ka-band satellite service delivering at least 12 Mbps to the remaining three, the report suggested.

It forecast up to 250,000 kilometres of access network and backhaul fibre deployed and up to 5,000 customer visits per workday during the eight-year rollout.

"Driving take-up should be the main priority for NBN Co over the coming decade," the authors wrote.

"Government should expect strong take-up of NBN services in the long-term if NBN Co sets prices to enable retailers to offer superior value to end users currently served by the copper network.

"Over time, we expect that Government would aspire to extend fibre even further," they wrote.

While NBN Co was to be a wholesale provider only, and could not set end user prices, the report noted that the wholesale price would likely be the largest input cost for a retailer.

The NBN implementation report recommended the establishment of "affordable, uniform, wholesale access prices" across the entire fibre network.

This would go against the traditional wholesale fixed-line model where exchanges in particular geographical 'bands' received uniform pricing, often meaning more expensive rural access.

However, uniform pricing would meet Government aspirations of delivering "affordable broadband to all Australians" by reducing retail price variability and drive uptake of NBN services, the report suggests.

"Fibre is the physical network platform of the future," the authors wrote.

"It has the potential to far exceed the short-term performance requirement of 100 Mbps ... if properly specified and deployed, the physical network should not need revisiting for 40 years and more."

6 comments in this discussion

"Is it just me or shouldn't the priority for this be the people outside of major population centres who have crap and beyond crap internet access. There are people out there who still can't access ..."

By crysis

Tags

nbn, conroy, fibre, kpmg, mckinsey, implementation study, wireless, government, broadband, coverage, nis

Exclusive: NBN Co withholds fibre upgrade price

Darwin prepares for cheaper broadband

How ISPs will tempt you in the NBN era

Wholesale talks stall Telstra, Optus NBN pricing

Breaking Stories

Bankwest credits IT for flat operating expenses

Conroy warns of cloud uncertainties

Hiring strategy vital to government IT transformation

Melbourne scientists ready 2013 bionic eye trial

ACCC declares Telstra wholesale pricing

Comments: 6

Graeme Harrison (prof at-symbol post.harvard.edu) May 7, 2010 8:44 PM	See my comments on the other article about the McKinsey NBN report re poor cost/benefit ratio. With a $4000/house connection cost, I really don't see the benefit in cabling whole streets of pensioners in small country towns. The incremental ROI will be incredibly poor on such decisions. And I suspect that "stretching the reach" to all the houses 2-20km from towns will have a similar issue, of costing (on average per the graph) $7,000 to install. If you run the fibre 3km from the nearest house (say) to a place that you know has no interest whatsoever in subscribing, who are you incurring that cost for? The whole ideology of doing the SAME thing everyone has flaws. Yes, do every population centre that has enough pre-subscribers, or business users... but I just can't see the problem with (in the first instance) only doing the main street of a town of 200. If in 10-20 years, there is demand to do a second phase, address that then. I suspect that if you have copper and wireless also available, the demand will not be there... as those older technologies will be pricing themselves at less than the NBN 20Mbps service. In short, I agree with subsidy of rural customers to some extent. But, it seems that if you pull the McKinsey report apart, you will find a really high ROI on connecting the larger population centres, and REALLY poor ROI on doing marginal tiny country towns and those living outside of such towns. I bet that extending the satellite service to those marginal users, or a mix of copper-based ADSL and/or satellite options, or even adding high-speed wireless in those towns MUST turn out to be cheaper, rather than going for near-universal access for reasons of ideology. Finally, there is the issue that satellite services are not just about bandwidth. In Sydney I have ADSL, but at my farm I have the existing Optus satellite offering at a "nominal" 512kbps. The problem in specifying satellite services in terms of ONLY kbps is that it ignores the higher latency inherent in satellite services. 512kbps with high latency is just enough for web viewing, but provides completely unsatisfactory call quality for VOIP. When any such higher-speed satellite service contracts are let, they should specify that two users on the network ought be able to have a reasonable Skype conversation over the network (ie specify an overall max latency for 95% of the packets, to enable common VOIP applications work with reasonable voice quality). Similarly, that same test should apply for a user on the satellite network talking to a fixed-line telephone service at the nearest capital city (though having both VOIP users on satellite should be the harder test). That way, some variables about how time-slicing is done, and packets are handled (at the satellite service end) would be changed, to allow voice to be 'easily' conducted over a satellite service that was 20-times faster than the current satellite service supplied to much of the bush.
umbria May 8, 2010 1:25 PM	Graeme, I agree that VoIP is critically important for rural Australians to eliminate their STD telephone costs, and this includes many pensioners who keep paying Telstra a monthly plan cost plus $1 per call because they don't understand that they have an alternative. The fact they are pensioners is irrelevant, however. The provision of fibre wherever possible to rural areas must be based on the comparative cost of the fibre ideal against the less satisfactory alternatives, not who the present occupants happen to be. Satellite will always suffer from latency due to the reality of the distance into space and back. As technology improves and with it people's expectations, this physical obstacle will leave satellite as the solution of last resort for two-way live communication, with wireless the second-worst. If we (the taxpayers) can lay fibre to rural addresses and still break even after fifteen years, then it seems a no-brainer to do so.
umbria May 8, 2010 3:11 PM	Geostationery orbit introduces an extra response-time delay of 0.4769 seconds (4 x 35,768 km / 300,000 km per second) if the terrestrial contact is directly below the satellite. This is in addition to any switching or relay delays that are probably comparable to those in fixed networks. For broadcast television this half-second delay is inconsequential, but satellite is a very poor substitute for two-way live communication, unless we can overcome light speed.
anonymous May 8, 2010 3:15 PM	Yes, umbria, well said. It's not clear which country towns Graeme has in mind. What is clear is that most rural and regional areas have the usual population mix, and the usual proportion of business users. Graeme talks of having a farm, so he should be aware that all farms except the hobby variety are businesses, and have at least the same comms needs as any other business. And good Net access for families is even more important in non-metro areas. Limited or non-existent other opportunities for accessing learning sources make it vital for these families to have equal Net capabilities with their urban cousins.
Graeme Harrison (prof at-symbol post.harvard.edu) May 8, 2010 6:34 PM	Umbria said "If we (the taxpayers) can lay fibre to rural addresses..." but in fact the rural addresses are that 7% which misses out. The wireless is intended to get those within reasonable distance from town, and the satellite for the remainder. I know about the inherent latency in satellite, but I was just suggesting that, in whatever satellite contracts are issued, they should specify 'as good as possible' performance under VOIP, rather than simply specifying a Mbps figure. The reason is that there are some variables, involving capacity, bursting and packet turnarounds at satellite end etc, where latency can be optimised. The reason I have VOIP at my farm, and have set it up for some older neighbours is because the copper connection is actually copper for 500m, then two separate 1970 radio links... and if the phone goes out in our valley, it can be out for days. And so the VOIP is an alternative for accessing emergency services or Rural Fire Service. I'm also aware of the cost argument that "it's cheaper to do the whole town while we're here" rather than just main street and others by demand (for small towns). But I feel certain that the NBN would have a 30% ROI if taken to every location where there are 10 or more employees in a non-farm business in Australia... and this drops to a 7% return if you try to get to 93% of the population, and it drops to a huge negative, if you try to get to 99% of the population. That's why there is the distinct 'knee' in the graph, as it gets more and more expensive as the distance between consumers climbs. So there is already a cut-off in the McKinsey study - all I'm saying is that if the ALP were not so ideologically driven, a commercial approach would be to set the cut-off where it ceases to make commercial sense.... Otherwise it is like building a school hall for a school with one pupil (the case cited in the stimulus spending). Go to some farmers and explain that the government can pay $7k in capital cost on his place, in total outgoings over the next few years, to spend all surpluses until Labor is voted out for over-spending (again). Then ask him if he wants it ALL spent on upgrading his existing working satellite internet connection, so he can watch more videos... or would he prefer (by then) a free 2-3kw solar system on his roof, to essentially wipe out his power bill for all time going forward. I'll bet most farmers would prefer to help get off coal. And in the urban areas, the high level of spend on trenching would only get my support if we put in multi-service general utility ducts. This would allow the concurrent undergrounding of all urban power lines (as 50Hz electricity supply electromagnetic radiation doesn't impact fibre signals). And I'd also like to provide for supply of non-drinking water in estates and medium-density suburbs in that same trench. This means that recycled stormwater etc would not have to be brought up to drinking standards, and houses could elect to tap into water that was able to be used for gardening, washing cars, washing clothes, toilets etc. Many farms (like mine) have riverwater for washing etc, but a separate supply from a rainwater tank for pure drinking water needs. The power utility companies would want the electricity cables to rise up from the ground at any termination (to avoid contact with water) but that is what is presently done with Telstra's 'street stump termination boxes' to avoid the 90v of the phone circuits getting wet. I'm certainly not suggesting that recycled water ought be done at the same time as the NBN, but only that the legal and physical arrangements provide that the trenches can be used for a variety of utilities. Currently in our cities, a roadway will be dug up to lay a new water pipe and a few weeks later that same roadway or footpath will be ripped up again to lay another utility. It ought be possible to use common trenching, especially in the denser areas. Getting rid of telegraph poles in our urban areas would remove a significant blight on the visual landscape. And if Optus were allowed into the NBN, then they would not bother also undergrounding their existing hybrid-coax cables hanging on those same poles beneath the electricity cables.
crysis May 10, 2010 9:23 AM	Is it just me or shouldn't the priority for this be the people outside of major population centres who have crap and beyond crap internet access. There are people out there who still can't access adsl as their phone lines are so old they struggle to handle dialup. The major metro areas are already oversupplied with internet access. Just for once, i'd like to see the rest of us get decent internet access. I am three blocks from being able to pick from three providers - i am stuck with flippin' Telstra as Optus and Vodaphone coverage doesnt come this far 'up the hill' (I am five minutes from a city with 15,000 people). My son lives on a farm 15 minutes from the nearest town and has wireless net only which only works 60% of the time - if he's lucky. The phone lines in the entire area are 1950's era lines and telstra won't update them. He can't get a satellite because dialup is available - except for when someone else on the same road is on the phone or when telstras own technicians physically knock the lines down lol. Yet again what we are going to have is the same areas of Australia being ignored, if you look carefully it's always % of the population, not % of the country covered. For those that will say that that is good enough - why should the same people always be missing out? You need the farmers - for your beef, your wool, your milk, your cereal crops, your fruit and vegetables, but they will be the vast majority of people ignored in the Semi-NBN. Farmers are not just centred on business needs - and the net is even more important to them for access to resources and for their families.