Back in the days when networks and networking services were designed, built and operated by telcos (telecoms companies to those of you who don't live in IT) there was a great vision of the future, where an information superhighway would connect everyone at high speed to networks and we'd be able to build fantastic new services. Then the web arrived and people said that would do, and oh how we laughed! The web was clearly never going to be any more than a machete-hacked information path through an information rainforest.

So by and large we ignored it for a while and got on with designing a real superhighway. And then soon after came the invention of ADSL, and suddenly we didn't have to put fibre everywhere because reasonably fast transmission could be done across existing copper networks.

The telco accountants and boards loved it, because suddenly there was no rush to invest in fibre any more. Instead of building an enormously fast network and then switching it on and hoping it would be filled, they could sit back and relax and watch while services used the slow ADSL network and let it gradually grow. No risk. But also no real gain.

History will show that the invention of the web and ADSL, happening as they did against the background of newly liberated telcos under strict regulation, were both brilliant inventions, but nevertheless set us back decades.

Were it not for them, we wouldn't have the web we have today running on a 'broadband' network that provides anything from 1 to 50Mb/s depending where you live. Instead, using older nomenclature, anything less than 2Mb/s would be called narrowband, 2-34Mb/s would be called mid-band, and only speeds above that would be classed as broadband.

Average data rates to the home would be 140Mb/s. At least that's what all our plans said in the early 90s, for rolling out from the late 90s onwards, and certainly intended for finishing by now. In fact, some of the projects I was involved in were assuming we would be providing 2Gbits/s to each home.

I even got a patent on a protocol I called Addressed Time Slices in 1987 that doesn't even start to be relevant until domestic delivery speeds exceed 625Mbits/s. Also in 1987, I worked out how to do 10Mbit/s to each mobile phone (uncontended), something we will only see coming reality in the next year or two as 4G rolls out (LTE and HSPA claim higher rates but only when there is no contention, and they expect 10Mbit/s to be a more reasonable expectation). In both cases, it will have taken over 25 years to develop stuff that could have been done in five.

Enough history, where can we go now? Will we get our 140Mbit/s soon, and then 2Gbit/s? Yes, we can do it, and global market forces will force it to happen, simple as that. The answer to the equally relevant question of whether anyone can make any money from providing it is also yes. Whether it will be telcos or someone else is a different matter.

We love wireless technology but few mobile solutions can deliver true Gbit speeds. Notable exceptions are optical wireless and ultra-wideband (UWB) and even these struggle. Both of these only work with very short range, unless extremely expensive kit is used to increase range.

But short range is OK sometimes, where there are enough people. High speed short range solutions could be used to make a sponge network in busy areas (a mesh net with so many links that it more resembles a sponge in its throughput characteristics than a mesh), and this could be an alternative to existing mobile providers' nets.

Possible implementations include jewellery nets, where tiny pieces of jewellery talk directly to each other and to those of other people and objects. A lapel pin may use optical communication or UWB to another belonging to a passer by.

Fleeting networks with links that sometimes may only last only a second or so could be set up. With a few possible connections between devices on the same people, multiple routes would exist for any packet to take even between two points, hence the sponge analogy.

But where things are fixed in place, fibre and local Ethernet or wireless  is always going to be a better solution in terms of cheap bandwidth. Although a home may need high total bandwidth, in most cases this would be divided among a range of appliances.

Fibre has enough bandwidth theoretically to carry millions of high definition TV channels, and it is hard (though not impossible) to imagine any services that will stretch its abilities.

The development of the world wide web has proved beyond any doubt that when faster bandwidth is available to people, the market develops accordingly, and new or enhanced services quickly fill it. Now, with high resolution cameras, video cameras, 3D HDTVs and so on, it would be easy to generate huge quantities of data to use up bandwidth.

Latent demand can be very high indeed. An average of 50MBit/s today latent demand would be a reasonable guess, with peak demand well into Gbit/s territory. That the telcos suggest that 'demand' isn't there is only an indication that markets lag capability, not proof of lack of demand.

People can only buy what is available to them, and also need time to understand the capabilities once they get it before they will fill it up. So companies should expect some delay between deploying a network and finding demand for it develop. Expecting it to happen the other way round is nonsensical.

The 50Mbit/s estimate for a typical household is based on a TV in the living room tuned into one channel while recording another, another TV in a child's bedroom, and two computers in use. Not terribly adventurous you may think. It would be very easy to exceed this a lot of the time.

Peak demand is based on the desire to be able to download a couple of HDTV movies (10GB each) to take out on portable players, and to be able to do so in a few seconds. 20GB is about 200Gbits after protocol overheads are added.

Doing the downloads in under a minute would take over 3Gbit/s. If a mid-range user were to do a backup onto a cloud service, with 500GB of data, doing so in under an hour would also need 1.5Gbit/s average. Compare that with the 340kbit/s available to  homes in many places today, it is 4400 times more! And the 500GB of files is just today's usage, tomorrow we will own far more.

Increasing use of the cloud for a wider range of services will quickly stimulate such latent demands. Latent demands turn into frustration if not fulfilled. Of course, we may never want to use the cloud to do backups on, but it is a good example of how there exist potential businesses that cant exist today because the net is far too slow.

Most fibre solutions use some sort of splitting, serving many homes locally, concentrating them onto single fibres upstream. Passive optical networks that could delivering gigabits to the home have been demonstrable since the late 1980s. It is really just price that is now limiting it, and price depends on many factors.

Global markets are required to provide enough volume to get component prices down, but business models also have to evolve. The model that was employed when ADSL arrived is severely flawed. It will still take many more years of letting markets gradually increase bandwidth according to proven demand before many attractive services are feasible.

At the moment, fibre is slowly being rolled out to certain areas where costs and benefits are expected to pay off, but a more adventurous approach would pay off.

A properly designed optical superhighway would do much better than a patched together information footpath of today. If we aim to put in 2Gbit's to each home, in both directions, then there will at first be enormous leftover capacity, but very quickly it will be filled.

With a network like that, anyone could set up web companies at home. The UK could become a global centre for information businesses. Lots of new business types would be stimulated. We would use the net far more to do socialising, gaming, and as our normal source of entertainment.

It would open up every area of media provision to strong competition. In short, it would fill up, and the UK economy would be highly stimulated. Business that became front runners on a high speed UK network would then hopefully flourish globally as other countries start to roll out similarly fast nets.

All in, the costs of deploying the network would be exceeded by the enormity of the benefits. The big question is who pays? During the 3G wireless spectrum auctions, the UK government reaped about the same amount in license fees from the telecom industry as it was estimated at the time to cost for a full UK fibre to the home roll-out.

If government were to invest a similar amount now, it would still pay for such a network and help enormously in the UK's recovery. Compared to the amounts wasted elsewhere, it is a manageable amount. And compared with trying to make the economy better via projects such as green economy, this would be a far safer bet with far richer rewards.

When it will happen depends on many factors. We have a government that demonstrates almost daily that it doesn't understand the information economy. We have companies with boards that are risk and investment averse. But internationally, pressures are building.

Convergence of media and its migration to the net may be very late, but it isn't never. It is happening, and announcements every week of new kinds of devices and new cloud offerings are accelerating the pressures causing latent demand to build up.

IT companies need fast nets to make their own devices more useful, and if they can't get access to existing nets that offer enough functionality, they'll build their own or at least bypass them with alternative solutions.

If traditional telcos won't act, there are lots of other companies who will be happy to grab their territory, and telcos are starting to realise this now. So I will expect some real action towards a real information superhighway to start in the next few years. And I for one can't wait.

http://www.futurizon.com