At The Register, Simon Sharwood sums it up:
New Zealand’s South Island, for example, is blessed with low ambient temperatures and lots of hydro power. A bit barn in such a location could have cost advantages – if you can get workloads there.
New Zealand’s poorly-served by submarine cables, which highlights the problem of latency. That’s in the too-hard basket for a while.
Southern Cross CEO Anthony Briscoe can think of many reasons why New Zealand doesn’t need another cable. He listed them to a full crowd at last week’s Tuanz After Five session in Auckland.
Briscoe also outlined the financial mountain he thinks Hawaiki must climb to make a profit.
Hawaiki CEO Remi Galasso wasn’t on hand to argue the opposing case. Instead Tuanz called on former Pacific Fibre CEO Mark Rushworth, now with Paymark. The short version of Rushworth’s argument was “competition is good”.
And it is. So is resilience. Briscoe admits there may be a case for another network to give New Zealand a back-up for Southern Cross.
In New Zealand, we tend to view resilience as a local problem. Having only one provider with two cables makes us feel insecure. In an age where the internet is the economy, a lot relies on one company and two cables.
It’s as if our economic well-being hangs on a couple of threads.
Single point of failure
Outsiders see the problem in reverse. The see one submarine cable as a single point of failure. Maybe one and a half points of failure if they take the Southern Cross two-cable design into account.
Either way, they see it a reason not to invest here.
It’s why, despite cheap, renewable energy, multinationals haven’t built New Zealand server farms.
The Tasman Global Access cable tips the balance. Add Hawaiki and, maybe, the proposed Bluesky cable. Now New Zealand looks a better prospect.
One, two… many cables
Almost overnight we will go from one company with two cables to three or four companies and four or five cables.
New Zealand is an attractive destination for applications where latency isn’t crucial. It’s ideal for large-scale secondary back-up. But also, maybe, for big data projects where there isn’t a great deal of data on the move.
We’re western. We’re stable. We’re part of the big trade agreements. Our laws are understandable and manageable, especially to lawyers in English-speaking countries.
We’re known as good people to do business with. Salaries are on the low side. We have skilled people, but even if we don’t have enough, skilled people are keen to come and live here.
Green energy calling
There is abundant, low-cost renewable energy. If, as expected, Tiwai Point closes, datacentre owners may find good deals. Bringing the big names to New Zealand would be a coup for a politician.
One issue needing thought is that submarine cables enter New Zealand in the north of the country. The spare hydro capacity lies in the south.
It’s easier and cheaper to move bits through fibre than electrical power through wires. You lose more electrons en route than bits.
It may make sense to put a new datacentre in Southland or Otago. Nobody cares how long it takes for electrons to move from Tiwai Point to Whangarei. Moving photons over the same distance adds another 25 milliseconds or so of latency in each direction.
More pipes, more traffic
We’ll leave those questions to the professionals. The key is if one or more big cloud operations open a datacentre in New Zealand, the traffic will boost profits for all the cable operators.
It would also change the economics of submarine cable links. At the moment the traffic levels are asymmetric. Far more comes down to NZ from overseas than travels the other way. ISPs and other cable customers have to buy symmetric capacity.
A New Zealand server farm run by a global operator would even the flow of traffic.
Of course a giant New Zealand server farm doesn’t need to be owned by an overseas multinational. This could be a good time for local entrepreneurs to move, by the time the new cables open for business it could be too late.
- International cloud providers reluctance to invest in New Zealand because of the ‘single point of failure’ speaks volumes. That alone is a case for more than one submarine cable operators. ↩