New Zealand - AustraliaNew Zealand’s internet centre of gravity is shifting.

Until now New Zealand data traffic was mainly east to and from the US. That’s changing. Increasingly our traffic goes west across the Tasman to Australia and Asia.

This will probably accelerate now cloud computing giants like Amazon and Microsoft offer data centres in Sydney and Melbourne.

The graph shows the trend is clear.

International cable trends
International cable trends – click on image to enlarge

Things have moved on since the last, 2013, data point.

Spark says last week the traffic was 56–57 percent trans-Pacific, and 43–44 percent trans-Tasman. Spark’s spokesman says if anything the trans-Tasman figure is lower than usual because of the holiday lull in work traffic.

Enter the TGA

The graph goes part way to explaining why Spark, Vodafone and Telstra plan to spend NZ$90 million building the Tasman Global Access cable.

The cable will run from Raglan to Sydney’s northern suburbs.

Alcatel-Lucent has the contract to build the submarine cable. Work starts soon. If it goes to plan, the TGA cable will be open by mid–2016.

That’s an interesting moment. Because if you extrapolate the graph data, it coincides with the time the two lines crossover.

Southern Cross holds the fort

Until then New Zealand will depend on the Southern Cross Cable Network: a single cable network operating two links; one heading west, the other east.

The existing network has served New Zealand well, but having all your economic eggs in one basket is not the best strategy.

Southern Cross argued there’s little need for more international capacity. It has upgraded network capacity as demand has increased. There’s plenty of headroom for the near future.

That’s needed. International data traffic out of New Zealand is increasing at about 60 percent year-on-year.

Southern Cross hits 100 Gbps

Today Southern Cross CEO Anthony Briscoe said in press statement that the network had recently completed live system trials with multiple vendors using 100 Gbps technology. This beats Southern Cross’ previous projections of potential capacity for the system

He says the network will support 12 Tbps capacity later this year up from the current potential of 7.8 Tbps and 14.4Tbps is achievable. That’s up from 240 Gbps when the network was first switched on.

Life in the old dog yet

Submarine cable networks have a use-by date but there’s plenty of life left in the Southern Cross network.

Despite the lack of competition here, New Zealand network customers pay the same price as Australians to use the Southern Cross network. Australia has competition between international cable networks. In effect, this means New Zealand already gets any savings we might get from competition.

And anyway, there’s something of a competitive question mark over TGA. Spark owns 50 percent of Southern Cross and around 45 percent of TGA. That means a race to the bottom on price is unlikely.

If it ain’t broken…

Why build a new cable if it isn’t needed for immediate added capacity and prices are already competitive?

Two main reasons:

First, New Zealand’s economy depends on international data connections.

The chances of both legs of the Southern Cross failing at once are slim. The network has a figure of eight design and the owners describe it as self-healing. That makes it reliable. There hasn’t been much downtime, but the chances of something going wrong are not infinitesimal.

Security, redundancy

Adding a third link boosts security. It means New Zealand is less likely to face catastrophic disruption.

Second, submarine cables don’t last forever. A new cable gives us more options for dealing with a replacement when the time comes.

The Southern Cross network started operating in 2000. Typically cable operations expect their networks to last 25 years at first, but if everything goes well they can then realistically extend the lifespan.

Southern Cross recently went through this exercise and now says the network is good until 2030. From day one the TGA will promise an extra decade or so of international links and it could be good until 2045.

An extra link also means there’s more capacity should circumstances change and the demand outstrips cable technology’s ability to deliver more. And adding capacity will give some service providers more confidence in either basing services here in New Zealand or providing services to this market.


If there’s a downside to the TGA project, it’s that the trans-Tasman link reduces the likelihood of a third player servicing New Zealand. It doesn’t help the economics of Hawaiki’s plans for a trans-Pacific cable.

2015 promises an interesting year for submarine cable companies with an official Netflix service arriving in New Zealand. The service accounts for about 35 percent of all downloads in the US. Even with caching, it will mean a jump in international data traffic. Southern Cross is confident it will cope with any surge in demand, even so, it looks like a good time to build an alternative cable.

CommsDay reports Alcatel-Lucent global fixed networks president Federico Guillen says his customers no longer see fibre-to-the-home as the broadband end goal. He says recent technological advancements mean the focus is shifted to vectoring.

There’s a context to Guillen’s comments. He was speaking at the CommsDay NBN: Rebooted conference in Sydney. The conference is looking at how Australia’s plans for a national broadband network will alter following this year’s change of government. The Coalition government has long been critical of the Australia Labor Party’s commitment to a fibre-to-the-premises (FTTP) network.

Guillen says: “Now it is about the [service] you provide to the enduser; it doesn’t matter if you provide FTTP, what matters is that you provide 50Mbps, 70Mbps or 100Mbps – that’s what the end user wants. And you need to do it more effectively”.

Alcatel-Lucent now talks of “fibre-to-the-most-economical point”.

It will be interesting to see if Alcatel-Lucent repeats this message in New Zealand. The company is a Telecom NZ network infrastructure partner,  it’s best-known project was the Telecom XT rollout which ran into teething troubles.

The company also works with Chorus and is involved in the Gigatown promotion. Alcatel-Lucent’s ng Connect open innovation program has commited to a $200k development fund for the winning Gigatown.

However the company missed out on New Zealand’s big FTTP project: Ultrafast Broadband. It’s rival  Ericsson won substantial UFB contracts from Chorus and from Whangarei’s Northpower. Meanwhile Huawei got the business from central North Island’s Ultrafast Fibre and Enable in Christchurch. Huawei also has some of Chorus’s RBI business.

With question marks now hanging over Chorus’s ability to finance the full UFB build and other UFB builders said to be facing similar financial problems, is anyone game enough to put a high profile Alcatel-Lucent speaker on a New Zealand stage to explain why building the FTTP UFB network might not be the best strategy?


Let your people go, just don’t let them go feral

Allison Cerra from Alcatel-Lucent talks about end-user trends at the Mind Storm conference in Auckland. The main point of her presentation is that technology is rapidly redefining the relationships between companies and their employees. This is unavoidable, so strap in for the ride.

While fibre is the future, it isn’t here yet. It’ll take another six or seven years to reach everyone. Telcos and ISPs could do far more with the copper already in the ground, but that may undermine the case for fibre.

When it arrives, the government-sponsored UFB fibre network will deliver download speeds of 100 Mbps and upload at 50 Mbps. The cost is little higher than today’s ADSL broadband plans. Fibre is a compelling product at a decent price.

If everything goes to schedule the last residential suburb will get UFB sometime in 2019. At that point the network will reach 75% of New Zealanders. Separate networks will connect rural New Zealand to the internet.

Why wait?

Business districts, schools and health centres are a priority, that means suburban homes – and the country’s many micro businesses are last in line. Some will wait seven years to get a fibre connection. That’s too long to wait for fast broadband.

And too long to wait when you consider UFB-like speeds are potentially available for many copper customers today thanks to VDSL technology.

ISPs selling UFB offer two speeds. Alongside the headline, premium 100 Mbps speeds they also offer 30 Mbps plans. VDSL can realistically offer 50 Mbps to most customers on the copper network. Alcatel-Lucent uses VDSL2 vectoring to do this in Belgium. In other words, engineers can squeeze UFB levels of performance out of the existing technology.

Not about the technology

So why aren’t we doing this in New Zealand? The simple answer is there are ISPs offering VDSL, but wholesale copper prices are regulated. ISPs pay Chorus a premium to deliver VDSL over the copper network. One could argue that premium is an artificial barrier erected to make fibre pricing look more attractive to users.

The argument against this premium is that VDSL is like a gateway drug. People will get hooked on VDSL’s higher speeds and will jump to the new network when it passes their gate.

The flip-side argument says VDSL is so good, customers will see no need to switch from it to fibre when the UFB network arrives. In this way it undermines the business case for a fibre network.

There’s a lot in this. There aren’t that many practical fibre applications at the moment for small businesses and residential users. Some observers worry the industry will have difficulty persuading customers to trade in their slower ADSL connections for fibre, let alone trade in VDSL. Slow sales of residential fibre accounts to date underline this fear.

Carrots and sticks…

If the premium charged for VDSL is simply to steer consumers to fibre, why not change the regulations so that the premium only applies once UFB passes a customer’s gate? This would have the added advantage of sending a clear signal about intentions.

If the premium isn’t about influencing customer behaviour, why have it? The Commerce Commission can make sure Chorus isn’t out-of-pocket by adjusting unbundled copper local loop (UCLL) unbundled bitstream access (UBA) prices accordingly.