Work has taken me away (quite far away) from blogging for some time now. There are a few subjects I want to blog about, particularly regarding the architecture of the UFB network. If time permits I’ll try to get to these in the next few weeks. Also, I might consolidate this blog with my other blog, since my posting has become so sparse.

Meanwhile, a quick post on the proposal for Chorus to be the UFB partner. In this article it is claimed that Telecom has 25,000 kms of fibre already in the ground: Telecom hopes for a break on regulation if it separates.

Really? They’re not including Southern Cross are they? That’s great, if it’s accurate. The problem is that probably more than half of this is running down state and local highways and linking towns and communities. It’s not the fibre in the street running past those homes in New Zealand’s urban areas that the UFB project is targeting.

To give you an idea of what is needed I’ll pick on Wellington, only because the council makes available the convenient fact that they maintain 654 kilometres of roads. This is good enough for us to make a first order assumption that the UFB project needs to deploy between 654 and 1,308 kilometres of cable and duct in Wellington to serve it’s 67,000 households. Why do I say 1,308 kilometres? Well sometimes we might need to run ducts down both sides of the street, in other cases there’ll be one duct down a street but lots of ducts drilled across the street to provide drops into homes. So simply doubling the route kms gives us a very nice fudge factor to cater to both situations.

Now if we use the National Broadband Map to check out Telecom’s fibre assets in Wellington’s southern and eastern suburbs we see that the fibre is actually quite sparse.

What’s more, this fibre may not be appropriately designed for FTTH, requiring some reworking. The reality is that the value this brings is not as great as Telecom might like us to think. It may seem like I am beating up on Telecom here, I’m not. As I’ve said before, having Chorus as a partner does makes sense, if done right. I don’t have the confidence that the bureaucracy springing up around UFB could get it right though.

This article (Supercity set task of coordinating broadband) reminded me of an old debate I had with someone on the funding for broadband.

Why is it that FTTH has had to face a far more stringent test of its business case than the billions of dollars poured into roads every year in this country? Even today there are deniers out there that still think the meagre $200 million per year over 8 years that will be spent on deploying FTTH is a waste. Yet anyone who looks at transport funding will know that road spending has almost no business case.

In this Auckland context, let’s take the proposed Waterview Connection, at $1.4 billion (though there are serious questions about that). You would expect that to spend $1.4 billion dollars would require a lot of analysis, first examining and setting the goals, and then looking at options on how best to achieve them. Well no, there have been absolutely zero comparitive studies according to NZTA. The decision has been made to build a road, and the only question has been what kind of form will that road take. So we can spend billions of dollars a year on roads this way, even while other countries are increasingly redirecting their transport spending away from roads. But when it comes to spending a comparitively meagre amount of money on something that all of the top countries in the OECD are doing (some have been doing it for quite some time now), and when it actually pays for itself, there are still so many naysayers. I would love to see a day when transport spending had to stand up to the same tests FTTH spending had in this country.

If there had to be a choice, between this one over-priced 4 km stretch of road with questionable returns, versus 75% of the country having FTTH, I think it’s pretty clear which is the better purchase.

Stephen Davies at Australian FTTH News has put up a recent presentation on the Point-to-point vs. PON debate. There’s some interesting debate in the comments too.

• Computerworld has a peice on Kordia’s submission. See Kordia warns of broadband risks (4 May 2009).
• According to Stuff, Vodafone is making a behind-the-scenes push to persuade the Government to establish a national consortium that would own telecommunications infrastructure. See Vodafone has ‘private plan’ (4 May 2009).
• REANNZ couldn’t find a suitable proposal from a potential supplier, so pulled the plug on its $15 million contribution towards a new submarine cable link. See Plug pulled on second Tasman cable (7 May 2009).
• Article at Stuff about financing the regional fibre companies, based largely on the submissions. See Debt a concern in fibre roll-out (8 May 2009).
• Some more on the failure to get a new Tasman cable underway. See Cable plan failure disappoints industry (10 May 2009).
• TUANZ blog post on Steven Joyce’s keynote at Telecommunications Day. See Joyce gives nod to fibre submissions (12 May 2009). The Minister is apparently saying LFCs will be allowed to wholesale layer 2 services now. But LFCs were already allowed to do this in the existing proposal. This is simply not enough, for residential users this must be mandatory.

It is worth commenting briefly on one small matter brought up by this Computerworld article about Vodafone’s submission: Vodafone expresses doubts about shape of broadband plan.

“The actual quantum of investment required is not yet clear. Based on the estimates of independent expert Dr Murray Milner, the real cost of the investment may be as high as $5 to $6 billion,” it says. “In Australia, the government has announced a total requirement of A$43 billion to reach 90% of homes which when mapped to the New Zealand context indicates a figure closer to $9 billion. There is clearly more work to be done on the business case.”

$43 billion is certainly the upper limit. More realistic estimates, based on real-world experience in deploying FTTH, both in Australian brownfields and greenfields, put the likely cost at less than half of this figure. So simply scaling that figure to fit New Zealand isn’t advisable.

Telstra is touting its HFC network: 100Mbps Telstra cable good enough for capital?

TelstraClear says it is investigating upgrading its InHome cable networks in Wellington and Christchurch so they can offer download speeds of 100 megabits per second.

The company says the upgrade would mean the networks could provide “substantially the same experience” as fibre-optic cable.

No, it really can’t. Not only isn’t this deployed yet, even if deployed it would only offer 100 to 200 Mbps of bandwidth shared across all customers in your neighbourhood (typically in the range of hundreds of homes). There is just no comparison to GE-PON and the now emerging and soon to be standardised 10GE-PON, providing 1 Gigabit and 10 Gigabits downstream shared between 32 customers respectively. Of course, you can expect 40 or 100 Gigabit Ethernet PON to follow too.

There seem to be serious fears circulating that only Telecom can build this FTTH network, and if Telecom isn’t chosen, it will make life very, very hard for whoever is. I’m sorry, but this is unreasonable. Telecom has made no such threats, and is not allowed to act anti-competitively, for example by pricing services below cost in FTTH areas. Besides which, no one, when in the process of spreading this FUD, has actually asked: is it in Telecom’s interests to do so?

A company deploying FTTH already has a strong incentive to connect homes in the process, and make migrating customers off copper as easy and affordable as possible, for both the retail providers and the customers. Any threat of Telecom only increases this incentive. FTTH is a long term investment, so anti-competitive behaviour of Telecom in the first 12 months can be anticipated and is not going to make a huge dent over the lifetime of the FTTH network.

Telecom might be able to orchestrate a short-term sabotage of uptake rates, but remember Telecom is obliged to provide wholesale services, and is now regulated to some degree. So don’t expect such behaviour to be a given, because it would end up costing Telecom. If Telecom started offering full line rate DSL to its subscribers, it would invariably find itself having to provide it wholesale. Not to mention, it would have to upgrade its network to suit this.

Ultimately, Telecom will have to decide how it is going to react. The best outcome for Telecom is to trade in its copper network, in some form, for a share in the returns of the FTTH network. The future of that copper network is scrap metal. Failing this, does it wind down and retire its copper network? Does it face losing customers and increasing its costs by insisting they stay on its copper? Does it run the network down and behave badly by offering below cost rates and poor service? Or, does it embrace the FTTH network, and let Xtra be free to do what other ISPs will do?

What will really make this end badly for customers, and good for Telecom, is if the Government is too timid, or fails to get its head around the real needs of users. Hence, don’t fear Telecom after the fact, fear what it (and Telstra too) say and do before the fact.

• Broadband is going to be more expensive, according to a report prepared for the Treasury by Telecom’s former CTO. See Computerworld: NZ’s broadband bill total could hit $7 billion and Low adoption could balloon broadband costs. (Relax, it’s not really a problem.)
• Telecom’s submission is all over the place. See NBR: Telecom to govt: let Chorus build a national fibre network. See Computerworld: Telecom joins the government’s fibre party. Some opinion, with added dose of fear: Opinion: Reading Telecom. See Stuff: Telecom’s broadband proposal.
• Not to be completely left out, some analysis of TelstraClear’s submission can also be found at Computerworld: TelstraClear: in or out?
• Vector and its allies in the Regional Fibre Group speaking, but not saying much, on Telecom’s submission over at NBR: Power line companies ally to push fibre, oppose Telecom.
• Orcon is expressing its concerns about the costs of cabinetisation and sub-loop unbundling. See Stuff: Unbundling “needs to be cheaper” – Orcon. See TUANZ: Orcon letter raises valid concerns over cabinet access. I am not entirely sympathetic to ISPs when they complain about cabinetisation and LLU. Because a) many of them seemed to treat cabinetisation like it was a surprise, which is stupid, and b) I always said LLU was a distraction and wasteful investment for New Zealand since we were coming to it so late in the game. It was clear (in my mind) that we should leapfrog straight to the top of the OECD with FTTH and that the ISPs were in the responsible position to push for this rather than LLU. I’m even less sympathetic when they think Telecom should provide it below cost if necessary. (Well, yeah, I’d like to buy everything below cost too.)

Telecom’s submission has been out for a couple of days now, the marketing machine has been in effect, and the news and opinion pieces have followed.

Telecom has made two proposals in its submission. The first, of course, asks the Government to invest in FTTH directly through Chorus, which would use this money to extend its existing fibre network, and which would then wholesale this dark fibre. The second, phenomenally, calls for the Government to forget about investing in fibre, asking instead for the Government to establish a new company that will only invest in a duct network. Telecom naturally believes the Government should outsource this to Chorus as well. Providers would then blow fibre through these ducts when they want to deploy fibre services to a premise.

It’s apparently not so obvious — judging by the media analysis — but this is a bad idea, a very bad idea. Telecom has a responsibility to its shareholders, so is naturally trying to protect its position, and the value of its existing copper network. This is exactly what its two proposals are about. They most certainly aren’t about providing bang for buck, or a smarter Government investment.

I explain in my submission why wholesaling dark fibre does not meet the needs of customers and providers. The drawbacks are too numerous to list again, but it boils down to this: in order to create a truly competitive and innovative environment for the residential and small business market, there must be a neutral, structurally separate, wholesale access infrastructure shared by all of the providers. In plain-speak, this means a suite of lit fibre services between the customers and providers, which the providers use to deliver their retail services over. These lit fibre services are an invisible part of the retail services offered by providers, operating at the most fundamental level. They are a common requirement for all providers, and aggregating this means that barriers to providers are kept as low as possible.

Any plan by the Government that does not understand this will ultimately be an expensive mistake, requiring future financial investments and regulation. Our neighbours in Australia and Singapore understand this. The majority of Government backed FTTH deployments around the world all understand this.

Note: wholesale dark fibre would still be used for business and other customers with special requirements.

So getting back to Telecom’s proposals. The proposal for a duct network proves the point I made in my submission. Wholesaling at the physical layer for the residential and small business market creates substantial barriers to providers. This benefits the large telcos, which are the only providers with the scale necessary to surmount the barriers, locking in large segments of the market in the process. Telecom has just decided to take it one step further and suggest the unbundling be at an an even lower layer. In turn raising these barriers even higher.

Never mind the increased cost, slowed deployment, reduced/slowed uptake, and so on, associated with this duct network proposal.

On the other hand, the first proposal does have some merit. There are good arguments that can be made for a single national network, and for Chorus to be a part of that. But only in the case where Telecom gifts Chorus and its assets in return for a share in a neutral, structurally separate company. There is no sound argument that this entity shouldn’t be structurally separate. Of course, as stated above, in any plan, this new company must provide an open access network for all residential and small business customers.

I have been a little occupied lately, but fortunately I still managed to complete a submission in time.

The proposal seems to believe that wholesale unbundled dark fibre is the best method to serve all customers, which is absolutely incorrect. The negative impacts identified in this submission are caused by this, and forcing LFCs to engineer the networks in such a way.

Unbundled dark fibre is only suitable for a small percentage of customers and is completely unsuitable for the remaining customers that make up the vast majority. It is important that this distinction is understood, and that the proposal targets the distinct needs of these two categories of customer.

The full 21 page submission is available in PDF format here.