Monthly Archives: February 2017

COMMENT: The real costs of FTTdp

Chastened by recognition that there is near zero political support for pervasive fibre to the home anymore, recent “fibre fanatic” campaigning has gone into promoting Fibre-To-The-Distribution Point or Fibre-to-the-Curb. Commentators such as Internet Australia and academic Mark Gregory say that FTTdp should replace planned FTTN installations at once on the grounds that it can support 100Mbps+ speeds, can be upgraded easily and that it only costs a little more than the FTTN it would replace.

But if something seems too good to be true then it probably is. Yesterday in the Australian, Gregory stated definitively that FTTdp would cost only $200 per home more than FTTN and that FTTdp could be upgraded to FTTH for less than $1,000 extra. Replacing all of the FTTN with FTTdp would only add $1.2-$1.5 billion to the cost of the NBN he opined, while further upgrades to FTTH could be absorbed within a further two to three years.

Sounds great. What a no-brainer! Except it is wrong. The definitive numbers, checked yesterday with NBN, are as follows: FTTH costs $4,400 based on real world numbers across a million lines, blended FTTN/B costs $2,200 based on the same scale sample. No one knows exactly what FTTdp will cost yet as it has yet to even complete a full-scale trial but best estimates at this stage suggest $2,800 per premises.

These are averages, distorted by the slightly inflated costs of FTTN in regional areas where there aren’t enough customers within sufficient radius to fill the 384 line capacity of nodes. In higher density cities, the cost of FTTN falls to $1,800. FTTdp, which pushes out fibre to within 30 to 200m of customers lacks similar scale advantages.

NOT TESTED: FTTdp hasn’t even been fully trialled yet. The experience with all other NBN technologies has been that the “on paper” costs swell in the field. It was the case for all five major technologies employed by the NBN to date and no doubt would be the same for a sixth one yet to be modelled outside of a spreadsheet.

But even in comparing a “modelled cost” with real observed costs, replacing FTTN would FTTdp across 6 million lines would more likely cost $3.6 billion in capital expenditure increases alone.

That’s not the sum of it. As the NBN cost benefit study of August 2014 clearly identified, much of the benefit from the multi-technology mix of the NBN comes from its speed to market. A benefit in the hand now is worth more than one in ten years time for a dollar spent now—a point which gets lost on NBN critics who seemingly care more about broadband in the year 2032 than the year 2017.

It takes NBN—or indeed any large broadband provider—about 12 to 18 months to complete the design, integration, IT, product and engineering requirements of a new network. The bulk of that work has already been completed for FTTN and that why that network is now being deployed at ever increasing scale. To bin all that and start again with an FTTdp topology which is only currently in trial status, would set back the NBN deployment by two years. So to replace FTTN with FTTdp across six million homes? Your real cost would not be $1.2 billion-$1.5 billion, but more likely $3.6 billion in additional capital cost and—assuming $42 monthly ARPU—about $1,000 of forgone revenue for every customer subject to a two year delay for connections. Six million customers delayed by two years means $6 billion of foregone revenue.

That is a $9.6 billion impact on the NBN business case. Hardly inconsiderable.

NBN is concentrating its planned 700K FTTdp line deployment in two specific segments: areas too sparse to be served by FTTN (because the economics of that technology depend on density and specifically whether there are 384 premises or less within a km of the node) and the Optus HFC network (where existing customers are used to 100Mbps headline plan claims and need to be upgraded to something superior for the NBN deployment to mean anything). The Optus HFC network requires considerably more node installation than the Telstra HFC network to reach NBN standards and its topology is ill-suited to an NBN network based on Telstra exchanges. FTTdp makes sense for these sporadic deployments.

MTM MORE FUTURE PROOF: The 2014 cost-benefit analysis of the NBN captured the overall issue quite well when it said “The MTM scenario leaves open more options for the future because it avoids high up‐front costs while still allowing the capture of benefits if, and when, they emerge. It is, in that sense, far more ‘future proof’ in economic terms: should future demand grow more slowly than expected, it avoids the high sunk costs of having deployed FTTP.”

“On the other hand, should future demand grow more rapidly than expected, the rapid deployment of the MTM scenario allows more of that growth to be secured early on, with the scope to then upgrade to ensure the network can support very high speeds once demand reaches those levels.”

While there is considerable noise advocating pervasive FTTdp, the arguments of these NBN antagonists fall apart under examination.

One argument is that FTTN cannot be upgraded. But copper lengths can be reduced and line speeds enhanced through use of 48 line micronodes, while the deployment of VPlus can increase normal FTTN speeds way beyond 100Mbps to 200Mbps or more. The technological path for VDSL is not dead.

Another argument is that in ten to fifteen years, all the investment in FTTN will have to be wasted. But FTTN will truly be a sunk cost by then and will be spinning off cash, benefiting from historical cost and speed to market advantages over the costlier, sluggish-to-build FTTH alternative. And, of course, the fibre to the node part of FTTN can be reused!

Then there is Gregory’s argument that you can just abolish the Connectivity Virtual Circuit and open the gates to maximum speeds. In fact, the whole idea of the CVC is to keep access fees for low end users cheaper: NBN uses it to offset about a third of access charging. Abolishing it would see a need to increase access charges by at least 50% to make up lost revenue. The best way to reduce the CVC is to suppress costs.

QUOTAS UNUSED: Gregory also argues that most NBN users are opting for small data quotas. Notwithstanding the fact that monthly usage is a different beast to speed (it reflects time online as much as the speed you have), average NBN households use 150Gb per month of data: most plans offer 500Gb or more. Typical users are not quota constrained. They simply have other things to do in their lives.

Then there is the argument that NBN needs to be “gigabit-ready” because you can buy gigabit plans in Singapore and Hong Kong. Again, complete rot. Gigabit plans in those cities are marketing gimmicks where, by very definition of what a broadband line can be used for, you will rarely exceed 5% use of your capacity and with compression advances unlikely to ever even half use in your lifetime. Not while TCP/IP remains a thing! Home gigabit is the broadband equivalent of a flashy sports car that cannot be driven legally at its top speed or George Michael or Andrew Ridgeley sticking shuttlecocks down their pants on stage. Nice to envy but not a genuine human right.

For an NBN that has to span millions of square kilometres of terrain, meet the market for legacy broadband services priced on a Telstra CAN valued at less than half the NBN cost and attempt to reduce risk to the taxpayer as much as possible, it shouldn’t even be thinking about gigabit. The object is to get the poor buggers numbering several millions off sub-10Mbps DSL tech that haven’t seen an upgrade in a decade.

The biggest lie of all is that we are all starved of speed, that proper broadband is some way off and we are all being denied our future because of Bill Morrow and Mitch Fifield. It’s what I, as an editor, identify as the “future tense” trick to deny the present reality. Between the Telstra and Optus HFC networks now and the existing live NBN FTTH/N network with over 2 million customers, there are several million people now able to access 50-100Mbps services. Nearly all medium to large businesses have access to fibre or wireless equivalents commercially (it is not in the NBN’s remit to service such customers). Have you noticed that sound of normality? The distinct lack of revolution? The sense that life continues on unchanged?

Indeed. Their lives have not been transformed by their new found superpowers. They are not clamouring for even more at state expense. Getting them to a gigabit is not and should not be a matter of policy focus.

Let’s worry about getting everyone else to 25-50 megabits at a price they can afford—both directly as customers and indirectly as taxpayers—for the genuinely deprived in what remains of this decade before we worry about 1000 megabits in decades hence.

That is simply good public policy. And that’s the mission of the NBN.

Grahame Lynch is the founder of CommsDay and a former global editorial director for America’s Network, Telecom Asia and Telecom China. He has reported and commentated in technology for 23 years