[LINK] NBN Wireless Service Looks Good: So do we need fibre?

[Richard Chirgwin]

Tom,

Yes.

RC

For the long answer, see Shannon-Hartley theorem plus various ACMA 
spectrum planning documents.

I understand what Beazley was talking about when he said "vomit point" 
(just when you think 'I'll vomit if I have to explain this one more 
time, that's when people start to understand it).

Thought experiment: wireless broadband for all Sydney with (say) three 
million households
Using the LTE data from slide 16 of this presentation:
<http://www.nbnco.com.au/publications-and-announcements/latest-announcements/doc/presentation-to-the-australian-information-industry-association>

10 users per cell with a good signal can expect approx 10 Mbps.

By that maths, and with the next technology, to serve 10 Mbps at peak 
periods you'd need 300,000 cells in Sydney.

Note that I'm not getting into arguments like the "wireless will 
overtake fibre" nostrum that's snared the ignorant among our 
politicians; nor am I anywhere near familiar enough with spectrum 
management to imagine whether you could set up 300k cells in Sydney with 
what's available; nor am I considering the potential real estate costs 
associated with 300,000 spectrum sites, nor the capital costs, nor 
whether it's worth pulling fibre to 300k cells without bothering to go 
the extra kilometre to connect the home directly...

(I have two mobile broadband devices, one 3 and one NextG. I'm not a 
wireless luddite.)

One of these days I'll find the Magic Wireless Fairy and whack it 
against the rock of physics until it's good and dead...

RC

Tom Worthington wrote:
> NBN Co have released white papers on their proposed fibre, wireless and 
> satellite products. The "Product Overview Wireless Access Services" 
> (August 2010) I found to be of most interest. This describes a fourth 
> generation wireless data service seamlessly integrated into NBN's 
> network: 
> <http://www.nbnco.com.au/content/upload/Product%20Overview%20Wireless%20-%20Final%2017082010%20v1%201%20%283%29.pdf>.
>
> The 25 page PDF document is 905 kbytes. Most of the 905 kbytes is due to 
> a photo on the cover: I guess if you a building a broadband network, you 
> can afford to waste bandwidth. ;-)
>
> The wireless overview is written from an unusual perspective, being a 
> description of the features provided by a non-existent service. The idea 
> seems to be that NBN is looking for suppliers (most likely of WiMax) for 
> the service they have envisaged.
>
> What NBN has described looks feasible and a desirable product, if it can 
> be provided at a reasonable price and at a level of reliability similar 
> to that of the existing wired telephone service.
>
> Those companies who invested in spectrum for wireless broadband, such as 
> Seven Network's Wireless Broadband Australia (WBA), might now see some 
> return on their investment, by selling the spectrum to NBN: 
> <http://www.sevencorporate.com.au/_uploads/Files/Energy%20Australia%20Signs%20Wireless%20Broadband%20Deal.pdf>.
> This would allow the Seven Network to concentrate on selling spare parts 
> for tractors out of a shipping container: 
> <http://blog.tomw.net.au/2010/02/transportable-data-centre-for.html>.
>
> However, if NBN can provide an affordable and reliable wireless service, 
> it might make their fibre offering unattractive for the average 
> consumer. It is likely that the average home owner will be using a 
> wireless link for the last 10m from the NBN's termination point in their 
> home to IT equipment. So the home-owner is unlikely to see much 
> difference in the service between a fibre option connection and a 
> wireless one.
>
>
>