A fibre-to-the-premise cost study prepared for the Treasury in February and released to Computerworld under the Official Information Act, says the cost of the government's planned broadband rollout could be as high as $10,000 per premise if only 25% of users adopt the network.
If 100% of those who can use the network do so, the cost would fall somewhere in the range of $2,950 to $3,400.
The study says that in order for a fibre-based infrastructure to be economical, it is essential there be a very high penetration of users.
It further warns that estimating costs for the deployment of a fibre-to-the-premise access architecture is fraught with difficulty, due to the many variables and assumptions involved in such an analysis.
Read also: NZ's broadband bill could hit $7 billion
Access architecture will connect premises with the fibre network now being planned by the government.
The study, prepared by former Telecom chief technology officer Murray Milner, now of Milner Consulting, says it will never be possible to determine a single value for the cost of deployment.
"Any figure quoted is at best an estimate based on some form of massive averaging and can only be used as a guide."
The costs for deployment would vary considerably by location, the report says.
"The big challenge in estimating the average values is our understanding of the distributions of the different variables. Unfortunately, there is little information available to help with any form of comprehensive statistical analysis."
The study investigates two different technologies for the deployment: passive optical networking (G-PON) and active ethernet over fibre (AEF). The latter's total infrastructure unit costs and investment are about 15% higher than that for G-PON.
AEF is today typically used for business premises, while the PON (G stands for gigabit) is best suited to the residential environment.
In a background section, the study says the estimate of costs for the deployment of a fibre-to-the-premise access architecture is fraught with difficulty, due to the many variables and assumptions involved in such an analysis.
"...it will never be possible to determine a single value for the cost of...deployment. Any figure quoted is at best an estimate based on some form of massive averaging and can only be used as a guide."
The costs for deployment would vary considerably by location. "The big challenge in estimating the average values is our understanding of the distributions of the different variables. Unfortunately, there is little information available to help with any form of comprehensive statistical analysis."
In a section on fibre deployment costs, the various options are analysed.
Aerial fibre deployment is never going to be universally available, the study says.
"Some local councils actively discourage the deployment of new aerial plant. Some councils, such as Auckland City Council have been very negative with respect to aerial deployment in the past but appear to be more receptive today. Hence, there is a great deal of variation in the potential for aerial deployment.
"No one has a perfect understanding of what the potential really is."
The study concludes that 30% to 40% of homes and businesses might be suitable for aerial deployment under ideal conditions.
The advantages and disadvantages of the other methods are discussed in detail.
A likely model for the distribution of deployment costs, developed by the study, is:
o 30% aerial deployment at $40 per metre;
o 30% low-cost burial deployment at $60 per metre;
o 30% standard burial deployment at $80 per metre;
o 10% open trenching deployment at $150 per metre.
The weighted average is $69 per metre.
That leads to a likely figure of $2012 per premise passed when averaged across all conditions.
The passive cost per premise connected (the cost of running fibre past the premises) based upon 100% take-up, sits in the range of $2950 to $3400. But if there is a low take-up, the cost increases dramatically. At only 25%, it could be as much as $10,000 per premise.
The study says for fibre to be economical, there has to be very high penetration of users.
"This is the basis for the argument for government to subsidise such a roll-out, as the demand risk over the first 5-10 years is too high for a publicly listed company to contemplate."
Video services are identified as important from a revenue and demand perspective. However, they are also a large part of the cost component because of the technical requirements.
The study notes that the take-up on G-PON infrastructure needs to be in the 75% range or better to get unit costs into a useful range.
"The return on investment will then be determined by the services supported and the time it takes to achieve this level of penetration."
Consumer costs are a significant proportion of the total investment required to deploy G-PON across the 75% of urban New Zealand.
Under one (high cost) scenario, consumers will contribute around 23% of the total investment.
"It would be possible to increase the proportion paid directly by consumers, if desired, by forcing them to also pay for the variable passive cost, or the cost of the premise lead-in. This would increase the consumer contribution to over 40% of the high-cost scenario," the study says.
"However, it is doubtful that this approach would lead to high levels of take-up, as consumers would be reluctant to pay this type of premium for fibre-based broadband services."
The government is seeking public responses to its proposal released on April 1, which contemplates a two-level structure, with a Crown Fibre Investment Company putting money into as many as 25 local fibre companies.