IN AN IDEAL WORLD most wireless links would whisper. WiFi does the reverse, bellowing out its messages, wasting most of the power it uses, and shouting down the neighbours. Cellular wireless links are more power efficient but they still have to raise their voices to reach distant base stations. You can draw the lesson from how we ourselves make conversation: whenever possible we get up close and talk quietly to avoid drowning out other people.
So the more base stations there are in a cellular network, the better. Transmission power in both directions can be reduced, saving energy, lengthening battery life and countering lingering health fears. Network capacity and data rates rise, user density can increase and optimal use is made of available spectrum. Everyone, on the face of it, wins.
In fact, reducing cell size in cellular networks looks like the only way to meet future demand for truly ubiquitous, fast mobile broadband as networks approach the theoretical limits of cramming data into radio signals. But installing new neighbourhood base stations is expensive, sites are hard to find, and each needs a costly fixed 'backhaul' link to the network backbone.
This is why many industry eyes will be watching the outcome of a deal announced this week between UK developer Ubiquisys and Japanese network operator Softbank Mobile which offers a model for putting a base station into virtually every building.
Swindon-based Ubiquisys will supply Softbank with femtocells, small base stations that are normally used to improve 3g or 4g coverage within a home or office and which use the subsbriber's own broadbank link for the backhaul. Vodafone already offers femtos for this purpose in Britain.
Softbank's plans for rolling out the femtos are doubly revolutionary. Firstly, the devices will be available for use by the public as well as by the host subscriber. The Ubiquisys femtos are smart enough to co-operate with neighbouring models to avoid contention and to cope with users passing between them. And they can be rolled out on an ad hoc basis with the potential of creating a high-density auto-configuring public cellular network.
Secondly, Softbank uses a business model almost guaranteed to ensure a big takeup - it is giving subscribers femtocells for free, bundled with free ADSL broadband.
This is critical because femtocells, though they offer huge advantages to operators, are at current prices unlikely to attract more than a niche market. How many people are willing to pay Vodafone UK's one-off charge of between £50 and £120, depending on tariff, for a private femtocell simply to improve coverage in their home or office? Not enough to qualify as a high-density femtocell network.
Softbank's primary aim is to win market share from rivals like DoCoMo, according to Andy Germano, vice-chair of the Femto Forum, the industry body set up to promote the technology. The cost of the ADSL line is offset by the savings on conventional backhaul links, which can account for 50 per cent of operating costs.
Softbank might be able to save money by deferring the cost of building or upgrading neighbourhood base stations, something of interest to any operator upgrading to 3G or 4G. It is also, says Germano, planning to flog services such as home surveillance.
Of course Softbank subscribers will still have to pay for calls but, remarkably, charges will ot be raised.
If the business model works, it could be copied elsewhere and kickstart femtocell use just as Freeserve helped take Internet use into the UK mainstream.
"We think this launch will encourage a lot of operators to look at their cost structure and that if they can offer [the system] for free they will," said Keith Day, VP of marketing at Ubiquisys.
His company has contributed some nifty technology in a device that Day says is sold at a fraction of the cost of existing femtocells. Each femtocell has to co-operate with its neighbours to avoid contention, as well as coping with public users passing between cells.
Day says they can also handle user bunching, such as when people waiting for a late bus near a femtocell all decide to ring home at once. The device is smart enough to hand the calls over to the neighbourhood base station so that it remains available to the host subscriber.
The impact on the subscriber's ADSL service is also minimal, Day says, because cellular calls typically use relatively little bandwidth and the femtocalls pack sophisticated load-balancing and quality-of-service features.
It is unclear whether the fixed infrastructure of a femtocell net uses less energy than a macro network of larger cells. But femtocells are definitely easier on client devices. "If you get a good signal the phone is forced into low-power mode which will extend battery life quite a bit," said Day.
So is Britain likely to get a Softbank-type system? "I don't know," Day said. "I think there is clearly a need. Vodafone UK has if nothing else demonstrated that femtocells are something consumers like. There is no reason given the cost equation why free femtos can't happen at some stage."
Actually, Britain already has a similar system but based on WiFi rather than 3G. BT's Fon scheme, under which subscribers can use each other's wireless routers, has been successful but hardly amounts to a homogeneous access system.
Femtocells have much further to go, but Day said, "The nice thing about femtos is that you don't need a big bang. You can start with very small numbers and build up. All you need to do is to provide the right incentive for consumers and businesses and the thing will grow."
The transmission power of a femto is around 200 times less than that of a WiFi device, according to ay, and data rates at close range can be comparable with 4G technology. So is there any chance that 4G links could supersede WiFi for links within the home or office?
"I can't see it happening," Day said. "I think WiFi will remain complementary." µ
For when you just can't take another long lunch break
Control your Android TV from an iOS device? Um, no
Somebody call the irony police
Agreement with the Royal Free NHS Trust doesn't give option to opt-out