Leaked NBN document helps the case for FttDP

In looking at options to overbuild the Optus HFC network, fibre to the distribution point emerges as a compromise that might soothe all but the most hardened theologist.

Leaked documents released yesterday confirmed what many telco observers have long known: That the Optus hybrid fibre-coaxial (HFC) network is not fit for purpose.

This knowledge was no secret; Optus wrote the network down by AU$1.4 billion in 2002, and said yesterday that it and the company charged with rolling out the National Broadband Network (NBN) knew that the network needed updating.

"Optus and NBN Co have always acknowledged that parts of the HFC network would need an upgrade to support the NBN's product set," an Optus spokeswoman said. "In advance of handover, there has been and continues to be major investment into the HFC network to manage subscriber growth and capacity demand."

Financially, the impact of needing to overbuild the Optus HFC is well and truly covered by NBN's contingency funding. This is exactly the sort of unforeseen event that contingency plans are set up to be used for, and, in the larger picture for the NBN company, it is an inconvenience to redo the technology to 470,000 premises, but it is hardly the end of the world.

For its part, NBN tried to downplay the significance of the document, and said it was developed as part of ongoing risk mitigation, and that the company regularly prepares for multiple scenarios in network deployment.

But that scenario provides some insight into the thinking within NBN.

Of the 470,000 premises that may need to be overbuilt, approximately 120,000 are close enough to Telstra's HFC network be to covered by extending it to them. For the remainder, the document looks at deploying one of NBN's FttX arsenal -- fibre to the premise (FttP), fibre to the distribution point (FttDP), and fibre to the node (FttN).

nbn-fttdp-optus-hfc.png
Costs per technology for replacing Optus HFC (Image: NBN)

It should come as no shock to learn that of that trio, FttP is the most expensive and would require a AU$600 million jump in peak funding. Even though FttP has the lowest running costs of all, at AU$125 million, the cost per premises of AU$4,400 results in a half a billion-dollar blowout in capital expenditure to cover a mere 350,000 premises.

Meanwhile, according to NBN's numbers, FttN turns out to be cheaper than using either the Optus or Telstra HFC network, and saves AU$50 million in capex, but costs more to run: AU$200 million compared to AU$125 million for FttP. Loss in revenue aside, a switch to FttN could save NBN AU$300 million.

Sitting in the middle of these two options is FttDP, which costs AU$2,700 per premise to deploy and has operational costs of AU$150 million. At around 60 percent cheaper upfront than FttP, and 20 percent extra for ongoing costs, it would take approximately eight years before the lifetime cost of FttDP would be more expensive than FttP.

For a company that is obsessed with deploying its network as fast and as cheaply as it can, FttDP could be a middle ground between price and performance.

It's a change foreseen by Shadow Communications Minister Jason Clare last month, who said he expects NBN to move from FttN to FttDP.

"NBN has also recently revealed that they are about to trial G.fast in the lab," he said at CommsDay Melbourne Congress. "I think it is likely that sometime between now and the next election, the new minister will announce that NBN will be rolling out fibre to the curb -- using G.fast."

Later that month at Senate Estimates, NBN CEO Bill Morrow said premises that are more than 1km from the node and unable to reach speeds of 25Mbps using FttN will be connected using FttDP.

"That's still fairly new, nascent technology only now starting to be deployed across the globe," Morrow said.

"What is new, senator, is the boxes that are just now commercially becoming available that are small enough to fit within the pit in the footpath, and that are reverse powered from the home to be able to make that possible."

In October, NBN also announced that it had conducted a trial of G.fast FttB technology, attaining throughput speeds of 800Mbps.

On copper runs of 100 metres from a basement to a fifth floor apartment of the multi-dwelling unit, speeds of 522Mbps down/78Mbps up were seen. However, NBN pointed out that during the trial, it has had to turn on VDSL masking in order to avoid interference with other VDSL lines; once the "full spectrum" is turned on, speeds should reach almost 800Mbps.

Results from NBN's National Test Facility have achieved throughput of 967Mbps on copper stretching 20 metres -- the typical distance from a residential lead-in to a street pit -- and 800Mbps on copper running 100m.

With FttDP able to get close to the top 1Gbps speed offered by the NBN for consumers on a short run, and at only 60 percent of the cost of going the whole hog and using FttP, it's clear to see why Clare thinks the future of the NBN could be FttDP.

In his October speech, Clare also admitted that should Labor return to power at the next election, it will be impossible to return the NBN to the approach used before the Coalition won government in 2013.

"I can't fix the mess this government has made with the flick of a switch or pull out every node or stop all the work NBN is currently doing without potentially causing more problems and wasting a lot of sunk investment," Clare said.

"If anyone thinks I can just click my fingers the day after the election and we can go back to the way it was, they will be disappointed."

The full FttP dream is long dead, that much is settled, but a pragmatic embrace of FttDP could be the next best thing.