For a long time the 'tyranny of latency' worked against New Zealand. From a data centre perspective, this is the delay an overseas customer experiences for data hosted in New Zealand. However, with new technology, this can be overcome and is now only a disadvantage in an increasingly reducing number of situations.
Competition
Who are we competing with in attracting the big companies to put up global scale data centres in New Zealand? Firstly, the US. Besides mostly being their home country, it is where a large proportion of their customers are. So investment in data centres is flowing into places where there are cheap electricity and business incentives such as the states of Iowa, Washington, North Carolina, and Oregon.
Outside the US, we are competing with many other countries who also want to attract global data centres. Taking Google as an example, the company has its European data centres in Finland, Belgium and Ireland. Similarly in Asia, Google's data centres are in relatively smaller countries- Hong Kong, Singapore and Taiwan. So country size may not be such a big disadvantage after all. Instead, Google's choices reflect the importance of international IP connectivity as well as business conditions, both the regulatory environment and electricity prices.
Enter TICS
Within this context of stiff competition, the killer blow could be TICS, the Telecommunications (Interception Capability and Security) Bill which passed its first reading under urgency in Parliament recently.
The Bill proposes to provide Government with unbounded discretion in the name of national security, law enforcement, and economic well-being. The Government will be able to require service providers to provide lawful interception, in essence a backdoor, into their systems. The Government has the discretion to require this from a specific service provider or a class of service providers. For example, it could impose the requirements on Skype alone or on all player-to-player messaging in online games.
It is highly unlikely that any major global online service provider will give the New Zealand Government a backdoor to their systems.
In addition, the Bill proposes to impose a 'duty to assist' which will require all service providers to decrypt the communications of their customers in response to a warrant issued to a surveillance agency. The overseas-based service providers are equally unlikely to hand over decrypted communications of their customers in real time if they have a choice not to do so.
Building or locating data centres in New Zealand provides exactly that jurisdiction and compulsion. Even if the Government doesn't actually impose these interception capability requirements, the fact that they can do so almost unilaterally at any time plus the duty to assist will be an unacceptable risk, particularly when the New Zealand customer base is miniscule.
This pushes the overseas companies into negotiating a deal with the Government before they consider New Zealand for their data centre. While the Government is clearly open to doing such deals, it sends an extremely negative signal for the way business is done in New Zealand by creating the need to do a deal in the first place.
International Internet cables
Assuming it goes ahead, the proposed "Tasman Global Access" cable between New Zealand and Australia will increase resilience and choice in international IP connectivity.
Pricing of international IP connectivity will remain as a big hurdle. Current pricing of international IP connectivity is 10-20 times that of data centres in competing countries. Except for the investors in the new cable, Telecom and Vodafone, it is unlikely that pricing will fall for the other Internet Service Providers (ISPs).
However, the standard pricing of international IP connectivity does not apply for big scale purchases, especially so when most of that traffic is going out of New Zealand. As the vast majority of today's international IP traffic is coming into New Zealand, traffic out of the country is like using empty highway lanes that have already been paid for. IP transit providers get the opportunity to use marginal pricing or other business models to offer lower prices to global data centres in New Zealand.
There is a second, often under-appreciated angle. International IP transit providers typically need to pay to receive traffic bound for New Zealand overseas. This pay-for-peering situation arises as the vast majority of the traffic is one way, into New Zealand. Global scale data centres sending huge amounts of traffic the other way acts to balance the traffic exchanged with peers overseas. This results in lower, or even zero, peering costs.
All three factors above- scale, outbound traffic, and peering costs- can combine to provide prices to global data centres that while still above Asian and European competitors, there are not by so much as to knock New Zealand completely out of the running.
Can we do it?
Clearly New Zealand has been completely unsuccessful in attracting the big companies to host here and build global scale data centres so far. Current trends, even taking into a boost from fibre broadband, will not make a fundamental difference. We need some sort of a step change.
One source for such a step change, not covered so far, is big data for science and research. So far that's been lagging in New Zealand compared to other countries. A step change would have come from the SKA (Square Kilometre Array) project had New Zealand got a decent share of the pie. We haven't, though I understand there is still some hope for small offshoot projects. Not really something we can bank on.
So it seems that the only realistic hope for a step change is attracting global online service providers to build and use global scale data centres. Is that just an unrealistic dream? Why would such companies come here when they can easily go anywhere in the world, including Australia?
On the back of a second international cable, the time has come for us to stop nurturing the dream. Let's go for it. We can do it if we set our minds and resources to it.
I'd like to see a task force of public and private interests, perhaps led by NZTE and backed by a cross-party political consensus, with a single focus- get one major online service provider to host in a global scale data centre in New Zealand. Do whatever it takes to make it happen.
Then we will truly be on the way to becoming a 21st century Aotearoa.
Vikram is an Internet analyst and commentator. He blogs at Internet Ganesha http://internetganesha.wordpress.com/. He is chief executive of Kim Dotcom's company Mega. These are his personal views and not necessarily those of his employer.
