Strawman: How to save the New Zealand economy
Put aside the idea that it’s Kim Dotcom who wants to build a
new cable connecting New Zealand with the outside world for a moment and think
about what we’re really talking about here.
Firstly, we’re talking about building a data centre. Nothing
unusual in that – we have many dotted around New Zealand, some large enough to
register on the international scale of such things. Between Orcon, Vocus
Communications and Weta FX’s donation of the New Zealand Supercomputer Centre,
we have several.
But this would be orders of magnitude larger – something that
would either power Dotcom’s new Me.ga service or cope with the demands placed
on Google, for example. It needs to be robust, it needs to be multiply
redundant and it needs power. Lots of power. Green power. Fortunately we have
that and even better, the Tiwai aluminium smelter is apparently going to be
coming available soon and it requires 610MW to function. That’s 14% of the
national output, which makes for a scary conversation with government whenever
the smelter’s owners talk about packing up and leaving.
Google’s combined data centres use 260MW as best I can
fathom which leaves us in a very good position to take over production of the whole
lot and do it entirely by green means. That’s quite important to a company like
Google, but don’t forget Facebook, Apple, Twitter (those tweets don’t weigh
much but by golly there are a lot of them) and all the rest. In fact this piece
in GigaOM nails it quite nicely so have a look at why North Carolina is the place these guys base their mega
centres. Hint: power’s cheap there – cheap but dirty (61% of their electricity
is from coal, 31% from nuclear).
How cheap? They pay between 5c and 6c per kilowatt hour,
which is a really good price.
According to Brian Fallow in the Herald Tiwai smelter pays 5c per kilowatt hour also, but don’t forget that’s New
Zealand money, not US money, so let’s call it 4c per kilowatt hour in American.
Clearly that’s a good price, plus it’s almost all green
which means a big gold sticker for any data centre using New Zealand.
So we’ve got electricity covered, if Kim gets his submarine fibre
built we tick off another huge problem. There’s not much we can do about the
latency between here and the US, so let’s ignore that thorny issue for now. We’re
conveniently located a long way from everyone so let’s move along.
There’s the issue of land which as we know is hideously
expensive in New Zealand. Unless it’s somewhere like Tiwai Point in which case
it’s not. I think we can tick that box off, particularly if you consider the US
pricing as your benchmark.
That leaves us with two major stumbling blocks. Firstly, the
staffing situation.
We need to produce enough graduates (or import enough
graduates) to staff this kind of monstrous facility and at the moment we’re not
doing that. We don’t have any push to get secondary school students into the
industry and we don’t have any long term plan to stop this incessant churning
out of management students and encourage kids into the world of ICT.
Without these kids coming through at all levels, we’re just
not going to get a data haven off the ground.
Because that’s what we’re talking about here – turning New
Zealand into a data haven where anyone can store data safe in the knowledge
that we treat bits as bits and that’s that. Nobody is going to trust us to look
after their data if we’re willing to send in the Armed Offenders Squad in a
chopper-fuelled moment of madness on the say so of some foreign national. It’s
just not viable.
The final problem then, is the legal situation.
We would need to become the neutral ground, the data
Switzerland if we’re to gain their trust. Publicly adhered to rules regarding
data collection and retention. Privacy built in, access only under the
strictest conditions.
But think of the upside – the PM talked about New Zealand becoming
a financial hub and while I get where he was coming from, that’s old school
stuff. Let’s become the home to all things data related instead. It turns our
long-time weaknesses (distance to market, isolation, relatively small size)
into strengths. Plus we’re New Zealand! Nobody’s going to invade us, we’re too
far away and too friendly.
Latency aside, what’s the downside of getting this done? If
we build the capacity we can attract a first mover in and if we have one, we
can attract more.
Customers from banks to insurance companies to individuals
to governments to movie studios (yes, you luddites, you) could make use of our
clean power, our isolation, our cheap land and our fantastic environment to
secure their precious bits and we would get a steady, reliable source of
revenue for the country that’s sustainable in all meanings of the word.
Have I missed anything? Why won’t this work? Is anyone
thinking about this?
Our electricity is remarkably cheap compared with the US or other places around the world that house data centres… and on top of that, it’s far greener. If you look at the example I linked to about US data centres they run on a mix of coal and nuclear power.
We’re ahead on points with our existing schemes: it wouldn’t be too outrageous to build more dams and provide even more power if we have the demand for it.
I don’t think the green argument stacks up. Marginal generation capacity is thermal. If the smelter goes offline & nothing replaces it, we’ll use less thermal generation than now. If a data centre replaces it, we’ll use the same thermal generation as now.
Latency is critical for some services (time sensitive ones, like voice calls, video calls, gaming etc) but less of an issue for cloud based services I suspect – and I had an interesting meeting this morning that I’ll write about for tomorrow on just this topic. Hopefully we can reduce it to being less of a show-stopper.
curses, I answered this and the reply vanished. Apologies.
It’s not the data centres themselves that I expect to generate the jobs but rather they’ll spawn more interest in the ICT sector, more development work to support the data centres and that in turn will lead to increased job opportunities. If we have a large data centre then we may well get developers based here as well (depending on the service of course).
I see them being more of a catalyst than a direct job creator.
See next reply for the latency issue.
"so let’s ignore that thorny issue for now. We’re conveniently located a long way from everyone so let’s move along."
usually when you say "lets ignore that for now" it means you are going come back and address it later. Since you don’t do this (and, in fact cannot since the speed of light is unchangeable), and latency is indeed criitcal to most data centres, then the rest of the conversation is largely academic.
In other words, if you cant solve the latnecy problem then it is irrelevant whether or not you can solve theother problems.
It might be a good place for off-site backups, where latency isn’t so much of a big deal. And there is a surprising amount of spare capacity on the Southern Cross cable at the moment, you just need to persuade the people sitting on it that it’s worth selling to you at a decent price.
Don’t expect this to generate a lot of jobs though. There will be construction work building/converting a building, but after that a good data center should be very labour efficient.
Hi John, thanks for that and yes, that’s a huge issue. But for some services I do think NZ would be a great destination, particularly with regard to the problems of power, land costs etc… We’re not the perfect solution but we’re a lot better positioned than some destinations and as the rest of the world realises how the US treats the .com domain and all that sail on her, I think they’ll start to look for alternative locations. We have to be ready for that.
The one problem with this is physics, things like the speed of light.
Any data centre in NZ providing international services faces challenges around latency. In the end it all comes down to the types of services you wish to run out of the data centre. Where latency is not an issue what you propose works. But where latency is an issue you want your data centre(s) to be close to the populations you serve. With Amazon’s EC2, for example, you can have servers in several zones to keep latency to your services to a minimum.
Latency even impacts providing services from NZ to Australia (especially with real-time type services). I know that a lot of us in the company I work for are waiting for the day that Amazon extends it’s services beyond CloudFront and Route 53 in Sydney. Having an EC2 zone in Sydney would significantly help us. (And provide significant pricing challenges to existing NZ and Aus data centres)
If we focused on providing services where latency is not an issue then the idea has some merit. If, for example, we had services like Amazon’s Glacier in NZ then we become the "Noah’s Ark" of global data.
But if you could change the laws of physics we could do more 🙂