Is Graeme Samuel inaccurate on the FTTN issue?

[Renai LeMayKeymaster]

hey everyone,

FYI with respect to this story:

Coalition NBN is ‘obsolete’: Samuel
http://www.afr.com/p/technology/coalition_nbn_is_obsolete_samuel_9NkLG77FSNeFmEdAbtqfaK

“The former head of the Australian Competition and Consumer Commission, Graeme Samuel, says the Coalition’s broadband plan uses obsolete technology and would cost billions more to deploy than claimed.”

I am currently contacting two major technology vendors and an analyst for comment on whether Samuel is correct, as Malcolm Turnbull has issued a statement flatly stating that Samuel is inaccurate on this matter. You can read Turnbull’s statement here:

http://delimiter.com.au/wp-content/uploads/2012/10/Response-to-Graeme-Samuel-23-10-12.pdf

I would rather get the objective truth on this rather than publish yet another “he said, she said” article with respect to the fibre to the node issue. Hopefully I will get some detailed responses on this issue today, and will be able to publish a detailed article looking at this situation tomorrow. In the meantime, feel free to discuss this issue and forward me any evidence-based information using this forum thread.

Cheers,

Renai

[Hubert CumberdaleParticipant]

I’m not sure what to say. Samuel is right and his reasoning is correct that should be apparent to most. I’m not going to go into the cost issues here but really the only reason one would object to calling FttN obsolete at this stage (2012 and beyond) is because of ones own bias. We’ve heard many, many times from various coalition members that the “NBN will be obsolete before it’s finished” etc, citing wireless technologies as a worthy replacement for fix line (the latest from Turnbull just yesterday! http://www.malcolmturnbull.com.au/media/transcripts/transcript-2gb-radio-22-oct-2012/) and now here they are trying to tell us that FttN is not obsolete? Really? They must take us for fools and quite frankly I am sick of the coalition’s special pleading and double standards here.

[Renai LeMayKeymaster]

hey HC,

do you have any technical evidence that Samuel is wrong? I’m looking at the technical angle of upgrading a FTTN network to a FTTH network.

Renai

[quinkParticipant]

The biggest problem with the upgrade from FTTN -> FTTH that I can see is not so much in the node.

If a household wants FTTH, they need to have fibre from the node to the household, and you may lay some dark fibre as well while you’re at it. If another household then wants fibre, there’s three options. Either you can lay a whole new fibre, use one of the dark fibres you’d have put in when you did another place or you can put in a passive splitter. When a third household wants it in that node’s area, you get to do the whole thing again. When a fourth household wants it, again. Fifth – again. The incremental cost is very high and decreases, but not until you’ve done a fair percentage of premises in that node’s area.

The only way to make FTTN -> FTTH a viable upgrade option is to at least build some fibre out from the node in advance in anticipation of that future upgrade. At which point in time, you’ll need to be asking whether it’d be more cost-effective to just do it.

And if you don’t do that, then you’d be less cost effective with FTTN if something like 20% or 30% randomly distributed in that node’s area want FTTH.

And right now, an argument can be made that 44% of people already want FTTH to an extent that FTTN can’t really serve as an alternative without making compromises. I think we’re already at a point where upgrading FTTN to FTTH to fibre for those 44% would be a bigger waste of money at a per premise per node basis than just doing FTTH.

FTTN -> FTTH is a good option if very people actually need FTTH, and if those people aren’t randomly distributed throughout a node’s area. Once you reach a certain percentage of FTTH users, FTTN is less cost-effective, and it’s very expensive to do a fibre roll-out one randomly distributed house at a time.

[Hubert CumberdaleParticipant]

eh? I don’t believe Samuel is wrong. Turnbull needs to provide the evidence that he is wrong if he believes that. If you are asking me for proof that he is right then that is a question you have to ask Samuel himself for the claims he is making but really it’s just an opinion based on facts, the fact that almost everyone agrees with him is probably what has Turnbull having a hissy fit today. But let’s looks at actual facts to humour him anyway:

FttP offers much faster speeds compared to FttN.
Projections indicate we will need much faster speeds in the future.
FttP can not only accommodate these needs but if anything faster is required can be upgraded easily.
FttN offers none of this thus has a limited life span.

Verdict: FttN is an obsolete technology due to the alternative (FttP) accommodating not only today’s needs but tomorrows needs as well.

[quinkParticipant]

In a nutshell, it means that for a particular area instead of NBN Co doing a giant job one area at a time and once, NBN Co needs to do hundreds of small jobs and dozens of medium-sized jobs over two, maybe even three decades.

It will likely cost £1000 to £1500, so probably $2000 or $3000 here in Australia, given all the concerns, but that cost would drop down over time as fibre inches closer on average to homes. That just kills it for homes, and even quite a few small businesses.

http://www.thinkbroadband.com/news/5307-openreach-reveals-fttp-on-demand-pilot-areas.html

Right now it appears to be free:

http://www.openreach.com/orpg/home/products/pricing/loadProductPriceDetails.do?data=NUCdC%2BIWJatTzHCYgGcar0msq1Tc5V%2FL3cp2E7tlxh9Z6rNZujnCs99NbIKJZPD9hXYmiijxH6wrCQm97GZMyQ%3D%3D

Probably because they don’t know what they should charge. And the upload speeds are pretty uncompetitive too.

And nothing about this addresses the basic problem that Telstra still has control of the copper, and will still have under a coalition government, which has all kinds of other implications.

[Tinman_auParticipant]

Samuel explains it in the AFR article when he says:

““Many thousands of nodes would be required . . . and those nodes are ultimately obsolete equipment when inevitably there is a requirement to take the fibre all the way to the premises.

“You can’t reuse [nodes] or remodel them or develop them to provide fibre to the premises and you basically just have to throw them out.””

It’s not that the technology it’s self is “obsolete”, he means it will be obsolete in the context of “If the network is upgraded to fibre end-to-end, the nodes will be no longer needed (i.e. Obsolete)”.

Mr Samuel is basically saying that with the requirements Australia will need in the near future, building a FttN and then “upgrading” it to FttP, half the old FttN system becomes obsolete.

If other words, he’s saying the Libs NBN is a false economy, and it’s much more efficient to build what we will need now without the waste of building a system that will be basically be discarded (the “obsolete parts being from, and including, the node to the connection box at the premise).

• This reply was modified 7 months, 4 weeks ago by  Tinman_au. Reason: I are a tag failure :(
• This reply was modified 7 months, 4 weeks ago by  Tinman_au.

[quinkParticipant]

> “You can’t reuse [nodes] or remodel them or develop them to provide fibre to the premises and you basically just have to throw them out.”

Depending on how it’s set up, you can reuse/remodel them. Maybe. There’s only one trial that seems to be publicised well, and that’s run by BT. It may work, and it’s more likely that it does than not. The only problem is cost and why anyone would put in nodes in the first place that there is an effective requirement for a certain percentage being FTTH to such an extent that you’d run fibre down most streets anyway in the next 10 years.

[Renai LeMayKeymaster]

OK, what I am hearing from industry sources-type people is that Samuel’s statement that the nodes can’t be re-used is incorrect these days. It used to be (circa 2007) that you could upgrade to P2P-fibre rollouts after an initial FTTN deployment, but not GPON (which is what the NBN uses). However, these days things are different — you can upgrade FTTN to GPON-style FTTH.

However, it also looks as though Turnbull is underestimating the risk in this approach. It’s not simple, it’s not 100% proven, and the costs may be larger than he’s estimating.

Thoughts?

[Hubert CumberdaleParticipant]

My thought is even so is it really worth the trouble? It is generally accepted that fibre is the future, if you are building a FttN network with FttH upgrades built in you are practically accepting that (bear with me) and you are accepting that FttN is or will be obsolete. The logic is that money is to be made from those wishing to upgrade to proper FttP, while filthy poors stay with FttN and filthy slow speeds (digital divide ftw!), eventually when the price comes down in their scenario everyone (or almost everyone) will have FttP. Just at a later date. Not very efficient no matter which way you look at it. Think of the labour costs and time wasted doing it this way.

[quinkParticipant]

Bunch of things…

Weird things might be happening in the UK re costs: http://www.ispreview.co.uk/index.php/2012/10/uk-government-sacks-man-who-questioned-bt-superfast-broadband-costs.html

Which in turn links through to this: http://www.ispreview.co.uk/index.php/2012/09/bduk-cast-questionable-eye-over-bts-uk-superfast-broadband-roll-out-costs.html

And all the way to Fujitsu: http://www.ispreview.co.uk/index.php/2012/09/uk-government-black-marks-fujitsu-superfast-broadband-bids-as-high-risk.html

Who also had problems in rolling out with NBN Co (albeit Fujitsu Australia, not Fujitsu UK). MT banged on about this some time ago.

Anyway, Stephen Davies of Opticomm has previously said that “[Every] NBNco greenfield community is deployed in this same manner using an Alcatel ISAM product with GPON blades, the same product which can take VDSL or ADSL blades.”

And that “[...] EVERY greenfield estate has a Roadshide cabinet (Node) with an OLT inside servicing that community.”

Here: http://www.malcolmturnbull.com.au/media/nbn-not-delivering-to-new-housing-estates/

The two of which I can’t really verify, but aren’t likely to be true for the majority of the NBN Co rollout anyway.

> However, it also looks as though Turnbull is underestimating the risk in this approach. It’s not simple, it’s not 100% proven, and the costs may be larger than he’s estimating.

I think most telling however is the switch from not running FTTP in FTTC areas and saying they have no intention to do so (November 2010):
http://www.ispreview.co.uk/story/2010/11/25/bt-uk-achieves-fttc-broadband-download-speed-of-70mbps-in-lab-tests.html

To trialling it with apartments in areas that have FTTP already (January 2012): http://www.btplc.com/news/articles/showarticle.cfm?articleid=%7Bbe4f74e9-e939-4aa8-b8a6-ac4a5140ecad%7D

To rolling it out across towns (October 2012): http://www.openreach.co.uk/orpg/home/products/super-fastfibreaccess/fibretothepremisesondemand/fttpod.do

After they get called this: http://news.techeye.net/internet/uk-worst-in-europe-for-superfast-ftth-broadband

I think that they are just doing the trial to see how much it will cost, because they have no idea. And that price point of 1000 GBP to 1500 GBP (all of which I mentioned above), as an incremental cost on top of VDSL, suggests that that means trench work until they have fibre running past more buildings.

Telkom (South Africa) is also rolling out MSANs:

http://mybroadband.co.za/news/telecoms/46644-1gbps-telkom-vdsl-and-ftth-infrastructure-roll-outs-begin.html

This place claims about 2500 Rand (but Internet in South Africa is crazy expensive and slow to begin with anyway): http://mybroadband.co.za/vb/showthread.php/435297-10Mbps-FTTH-takes-on-Telkom-ADSL

—-

In any case… in BT’s case, it looks like it started in 2005 (and MSANs from Huawei and Fujitsu, http://en.wikipedia.org/wiki/BT_21CN ) there was a trial in 2007, and people on them complained about crappy DSL, but only in 2011 did things move forward enough that they actually tested things like alarm systems over FTTP:

http://www.bsia.co.uk/21cn-communications-project

—-

As a side node, I found this from BSNL (India), in 2009:

http://www.ofcnfoec.org/osa.ofc/media/Default/PDF/2009/09-Yadava.pdf

Where they say VDSL OPEX is > 1.4 x PON OPEX, cost of repair is 1.6x and power consumption is 20x. Probably not right, but pretty impressive :P

—-

Here’s a 2007 comment from Mark Newton, and it’s still as relevant as it is today:

http://forums.whirlpool.net.au/archive/876625#r13910111

> So the fibre network used to build FTTN will only be useful for FTTP if it’s installed with enough cores to connect each house in the neighbourhood serviced by the pre-FTTP node.

> Which is isn’t. It never is. Nobody’s pulling high-density cables into RIM cabinets. Usually 12 cores at most.

> If someone is going to contrast FTTN against FTTP/FTTH, it’s important that they understand that the technical and economic differences between them mean that there’s no upgrade path from one to the other. This notion that FTTN is a “stepping stone” to something else is pure fantasy. If an FTTN network is built you’d better like it, because it’ll be around for a long, long time to come.”

And MT still has this on his site and hasn’t retracted it or anything:

http://www.malcolmturnbull.com.au/faqs/frequently-asked-questions-on-the-nbn/

> understanding that FTTN or FTTC costs only a third to a half as much as FTTP, but up to half of this outlay will have to be written off if there is a later switch to FTTP

—-

In any case. None of the FTTN -> FTTH stuff can work around the fact that an piecewise upgrade to FTTH means constant capital works at large upfront expense to customers over a period of twenty years, nor the problems with maintaining copper, nor can it work around the fact that a Telstra tax has to be paid on all that copper to maintain it.

In a nutshell: MSAN + FTTN = massive capital works in laying fibre every time some random person wants to upgrade to FTTH, costing thousands. It might cost hundreds once most people in an area have it and fibre runs along every street, but at that point you’re wasting money on the node.

Build FTTH and be done with it.

[Tinman_auParticipant]

Wow, interesting Renai. And on a few levels.

This could mean that MT is kind of “back-dooring” his parties stated stance on the whole NBN thing by making it sound different enough to Labors plan that the Libs will go for it, but then later on he can turn around and ninja-fit actual FttP…

So what level of risk is there in it actually? And what sort of costs? It’s a real shame they wont actually release something that can be compared and debated…if they have a plan that would effectively end up with Australia with the same network (FttP)as Labors, but they really can get “stage 1″ in “faster, cheap, better” then I’d probably view their plan a lot more favourably.

The Libs themselves are their own worst enemy a lot of the time…

[quinkParticipant]

The fibre needs to be put in to about 60% of premises in the next 20 years pretty much no matter what, unless we’re willing to compromise on stuff a lot, which will affect a lot of the things we’re getting faster broadband in the first place.

It’s better to just get FTTH that rather than run two networks.

Although, now that I think of it… isn’t the suspicion that the coalition won’t roll out FTTN in HFC areas?

What happens to this FTTH plan in HFC areas?

[quinkParticipant]

And I think the ultimately most telling sign of this is that we heard no mention of this from Malcolm Turnbull until this month. It smells like he knew his plan wasn’t going to make it the election so thinks that with this addition it just might succeed. Problem is, if he can just get one or two sentences to the electorate at large, they’ll swallow it. The drawbacks are that explaining the risks, additional costs and compromises implied by this will take a bit longer to explain, and the only way to refute it is with a statement like “If we know we need fibre passing, even if not connecting, the majority of premises very soon, as your plan still implies, why would we do capital works a hundred times in an area over a decade or two instead of just once?”

[quinkParticipant]

> Technology Futures Inc is a US advisory that sets benchmarks for telecommunication assets – the kind of data that fund managers use to put a value against a carrier like Telstra. It sets the depreciation life of fibre at 25 years, meaning today’s NBN build will only “expire” as an asset 2037. Copper, on the other hand, has a depreciation life of just 15 years – which means much of Telstra’s network has already passed its asset lifetime.

http://www.abc.net.au/technology/articles/2012/07/26/3554003.htm

Why would we maintain copper when it lasts for 15 years when we could instead put in fibre over about 10 years and have that last for 25 years. Sure, that 15 year number is a bit low in comparison to the likely lifetime, but no more relatively lower than that 25 year number.

[Author]

Posts