rstuart there are very many frequencies that aren't visible and simple line of sight is all that is required, especially if there is a diverity of routers. It's just like fibre, but without the fibre and I'm sure the next few years will see fantastic development. They're already seeing 500MBytes/sec (yes, bytes) they say and that's pre-first generation. Back in 99 when you had a V90 modem did you seriously think that we'd be discussing this? why do you find it so hard to imagine the next logical step could be as huge and as rapid?

Let's face it, if fibre was a commercially feasible project in its own right it would not need to be done this way - as a shonky monopoly that has to rely on non-compete clauses to even pretend to be viable. Someone would have snapped it up, so why didn't they?

Remember, Telstra and Optus have 20 years of non-compete, do you reckon they'll be sitting idle and leaving all that infrastructure as 3g/LTE? What about the other players? What about new ones, commercial and otherwise who want to set up local wireless access points using the one connection to the NBN, in a sort of adhoc implementation of wireless distribution? I reckon I could be happy with a 100Mbps feed shared with 7 others via a decent router and there'd be no reason users couldn't rout on to their own wireless subnet. But the NBN just lost 7 customers out of a potential 8. Telstra/optus etc still get their mobile service dollars though. Lose/win for NBN vs the telcos.

Here's a nice little piece

http://www.news.com.au/technology/what-the-hells-a-femtocell-in-short-it-fixes-mobile-coverage-blackspots-in-your-home/story-e6frfro0-1226104197035

Mr Williams said that while the femtocell device solved mobile phone reception problems in the home, it was not an alternative to fixing poor coverage areas.

"Over the next year we're deploying 700 base stations across Australia and in the past three years or so we've invested $1.6 billion in infrastructure and installed 660 odd base stations," he said.

Telecommunications analyst Foad Fadaghi told news.com.au Australians should be aware that any femtocells sold now may need upgrades in the future.

"The terms of the deal are 24 months, and over the course of the next 24 months most of the carriers will be launching LTE or next generation 4G networks," he said.

"Consumers have to be conscious that they might need to upgrade these devices down the track as well, think about that in the context of going into 24 month contracts."

So even within 24 months Optus is forecasting this generation of the femtocell technology to be obsolete, because they're also talking 4g and did you see that data allowance?

All this and the NBN has barely got past trials.

@Shadow Minister: As per this article, there is no application other than live HD TV that requires more than 1Mb/s.

Then I disagree with the article. Pictures on web pages are often over 1MB(ytes). Flash on web pages are often over 1MB. Doing a google image search on "barbie" downloaded over 1MB on my link. Displaying my wife's facebook page downloaded over 1MB. Each of these things happened relatively quickly on my link, as within a second or so. That's comfortable. Waiting 10 to 12 seconds for a web page to display is not.

You may say it doesn't matter. But it does. That under 1 second wait as opposed to a 10 to 12 second wait is what makes web applications like Microsoft Live Mail, GMail, and Google Apps, Google search suggestions and so on usable. If they start pausing or connections become intermittent they become much less attractive.

Let me put it another way Shadow. The hardware advances you electronic engineers consistently made over a 30 year period are nothing short of a miracle. You've outrun every other profession except one, but they are the only one that matters. The software guys have been able to use everything you have given them, and still want more.

@Shadow Minister: This can be resolved by spending 10% of the NBN's budget, and for $4bn upgrade the average to close to 8Mb/s.

Indeed. And both the Liberal and Labor governments tried to do that. What you keep ignoring is neither managed to make it happen. The incumbent Telco's were too intransigent. You keep banging about the technical side as though that the only consideration that matters. It isn't. We ended up where we are largely because of logistical and business considerations, not technical.

Where we are isn't so bad. It's probably better for the tax payer in the long term than paying the incumbents to upgrade their networks. The downside is the increased risk. One part of that risk, takeup, has been neatly sidestepped assuming they keep the price the same. That leaves building it on budget - fingers crossed.

@Antiseptic: why do you find it so hard to imagine the next logical step could be as huge and as rapid?

May be because using visible light as your carrier isn't a logical step for outdoor transmission, given there is the ruddy great noise generator called the Sun in the sky? Maybe because there are a whole pile of reasons nobody goes much above 5GHz for anything other than point to point?

@Antiseptic: most of the carriers will be launching LTE or next generation 4G networks,

It was you I went through 4G figures with on this thread, right? Or are we just have this discussion pretending those figures don't exist?

Let me spell out the conclusions just in case you've forgotten already. In order to supply what the NBN is supply _now_, we would have to do a few things:

- Roll out 3.9G.
- Up the spectrum available to the telco's by a factor of 10.
- Design, and manufacture carrier grade picocells capable lasting for decades when installed outdoors on a pole (current ones are indoor only, because no telco in their right mind would deploy something reaching so few customers).
- Come up with a 16 element phase array antenna that is actually capable of surviving on a pole.
- Install said picocells and antenna at roughly 200m centres across the country.
- Get everyone to install a phase array antenna on their roof.

Or of course they could just use fibre we are running down the street to service those picocells anyway, which is potentially 1000's of times faster and more reliable.

@Antiseptic: did you see that data allowance

Err Anti, that 500Gb allowance was on his land line. The femtocell uses his land line to send data back to Optus. He was saying he was in the fortunate position of not having to care how much data the femtocell used, because of the prodigious data allowance on his land line. Its not so prodigious really. TPG sells unlimited data for $60/mo, although at a horrible contention ratio that makes it unusable for a femtocell.

rstuart, light has been used as a medium for outdoor transmission since Roman times and probably before. They used heliostats and signal fires, but I think the technology might have moved on a little.

Light, especially coherent light, propagates well in air and scattering is only a problem over great distances. The tech that is being tested doesn't even do any modulation of the carrier except switching it on and off. In radio terms it's straight CW, just like morse code on shortwave in the early days of radio. How far do you reckon you'd be able to see a blinking light?

rstuart, the point about the data allowance is that even for the chap in question, it was way bigger than he needed. So much bigger than an extra bit of traffic was not worth worrying about. That was on current tech. Why on earth do we need 1000 times more bandwidth when we can't use what we've already got? That is the question you keep dodging and it's the central one.

It's precisely that sort of blinkered approach that has saddled us with the white elephant that will be the NBN.

Your argument comes down to "it's happening, so there" and a reliance on the capacity of 3g for comparison and even there the NBN has trouble holding its own on the basis of genuine need.

As I said, do you also advocate a 2000 lane National Highway Network? Look at how much traffic it can carry! Think of the possibilities!

Let's make a little wager. I'll bet you that in 10 years you'll be using a wireless device for 50% of your internet activity and probably more, since it will be the cheapest option. I'll also bet that you'll not have any genuine use for the massive pipe into your home and it could be easily replaced with a 4g service without you noticing. That's even if the NBN has actually managed to go past your door by that stage.

BTW, the NBN isn't supplying much "now" at all. It will in a few years.

Here's another piece of new tech that makes the NBN look sick using a distributed model, much as I suggested earlier.

http://www.theaustralian.com.au/national-affairs/wireless-challenge-to-future-proof-nbn/story-fn59niix-1226104623595

In a paper released yesterday, Mr Perlman claims his new wireless technology breaks Shannon's Law, a fundamental theorem of communications that posits all users on a wireless network must share bandwidth.

This means that, on conventional wireless networks, download and upload speeds fall as more users are added to the network. But with DIDO that principle is turned on its head.

"We know we can get to 100-fold what today's cellular systems provide, and we are optimistic we can get to 1000-fold," Mr Perlman said in a recent talk at Columbia University, where he first publicly described the DIDO system.

His paper states: "The potential of DIDO is to have unlimited numbers of simultaneous users, all streaming high-definition video, utilising the same spectrum that a single user would use with conventional wireless technology, with no degradation in performance, no dead zones, no interference between users and no reduction in data rate as more users are added.""

The more I see, the more it looks like you're just not up to speed with developments.

There were always 2 types of engineers: those who see problems as challenges for the creation of something genuinely new and those who "know what can't be done". In geotechnical engineering, the latter build bridges for Main Roads using standard designs, while the former work out how to make buildings stand up in places where buildings "can't be built". Unfortunately, the NBN is the product of the Main Roads types.

I'm beginning to grasp the problem I have in discussing a vision that isn't written down in a standard spec. You "know what can't be done"...