Vector, Northpower lab-trial 1000Mbit/s fibre

Following the lead of the NBN across the Tasman, lines companies Vector and Northpower are about to trial super-fast fibre. Intriguingly, the Commerce Commission is not a fan of the G-PON technology involved (keep reading).

Old schoolers - and that includes some in Treasury - think copper-based DSL will provide all the bandwidth we need for years to come. On a good day, ADSL2+ can deliver around 15Mbit/s; VDSL2+ - if you live literally on the doorstep of an exchange, can multiply that (unlike fibre, copper-delivered broadband degrades rapidly with distance).

A fibre-based connection of up to 100Mbit/s (in both directions) is complete overkill, say the naysayers.

Cheerfully ignoring this argument, Australia’s National Broadband Network Co recently announced its intention to roll out not 100Mbit/s fibre, but provision so-called Gigabit (1000Mbit/s or 1Gbit/s) connections.

For those who remember 56Kbit/s modems (or, heck, 2.4Kbit/s) that's 1,000,000Kbit/s.

Here, Crown fibre contenders Vector (pitching for the UFB contract in Auckland) and NorthPower (Whangarei) are about to begin their own Gigabit fibre trial.

Crown Fibre Holdings plans to begin real-world technology trials later this year, working with councils, UFB respondents and others to gauge the practicality of new fibre deployment techniques such as shallow trenching - the better to discover cost-effective methods for the big roll-out to come.

The Vector and Northpower trials, on the other hand, will be purely “lab-based” a spokeswoman told NBR, and centre a technical assessment of G-PON (Gigabit Passive Optical Network) technology from four different manufacturers.

The trials are due to begin before the end of this month, and last for around three months.

“Some are still pinning their hopes on copper, yet that is fast reaching capacity,” Vector chief executive Simon Mackenzie said in announcing the trial.

But the power line pair also emphasise the new 1000Mbit/s focus isn’t just about raw speed.

G-PON, which is already used in the US and Singapore (and the NBN’s choice in Australia) is lower cost to deploy because a single end-point can serve up to 32 homes (compared to more expensive point-to-point systems).

NBN chief executive Mike Quigley has said that the move to 1000Mbit/s will not increase the Aussie project’s budget (which, until Saturday’s election at least, remains $A43 billion - $A26 billion of which will be from the Crown).

G-PON also uses less energy, is only at the beginning of its lifecycle, and has been successfully integrated in fibre networks by Vector and Northpower for several years, the companies say.

A visitor from Mars could reasonably assume - even their own experience already with the technology, and others’ overseas - that G-PON is already pretty proven. But the trial won’t hurt as various UFB contenders jockey for public and political attention, and parade their skills, ahead of Crown Fibre Holdings’ October decision.

The Telecommunications' Commissioner's objection
On paper, G-PON looks perfectly spiffy from a technology perspective, and an excellent way to efficiently reach the mass market.

But the Commerce Commission has a political objection, which is where things get intriguing.

Although he has no formal role in the UFB process (Labour wants him to), Telecommunications Commissioner Ross Patterson has had a lot to say over the past few months, on everything from layer two services to regulating bundled pay TV services.

At the Tuanz Telecommunications Day conference in Wellington earlier this year, Dr Patterson told an industry audience that regulators in Japan had run into practical problems with point-to-multipoint, passive network architecture (of which G-PON is an example).

Such a set-up could not be easily unbundled, the commissioner said (see page 17 of his presentation online here; PDF), if the government wanted to open a certain section of a fibre network to competition later on. And a fibre, once laid in a set multipoint configuration, could be in place for up to 50 years.

Point-to-point technology, on the other hand, was much easier to slice and dice.

Dr Patterson also disputed that point-to-point was much more expensive than multipoint technology, saying a once large differential had narrowed to 12%, based on recent European examples.

For a regulator, that’s just a snick. For Crown fibre contenders juggling their sums, it could look larger.