IT's role in saving the world - and your wallet

As the weather changes, the temperature climbs and the climatologists start buying houses on mountaintops, the message has got through: we have a problem. Oddly for an industry that prides itself on selling solutions, IT has been slow to identify its role in helping bail out the planet's overheating. That's changing. Groups are starting to find fertile ground for cross-pollination between power generation and information technology: smarter is cleaner.

Cutting greenhouse gas emissions is essential, but the search for cleaner power sources is only one part of the process. Even if we came up with the perfect power source tomorrow with zero emissions, no long-term waste and no requirement for fossil input, our ever-increasing outpouring of energy into the ecosystem would have long-term repercussions.

Efficient use of what we generate is of equal, if not greater, merit than cleaning up our sources - according to the Commonwealth Scientific and Industrial Research Organisation in Australia, 70 percent of the fuel used to generate electricity is wasted in transmission losses before it reaches its destination. Of the 30 percent that actually arrives, half is lost due to inefficiencies in homes and offices. If we're losing 85 percent of the power we generate before we use it, fixing the generators is not priority number one. And this is an area where IT not only has a role to play, it may be the only option that makes a significant difference.

To use less energy, we must turn stuff off -- or turn it on when there's spare capacity, so allowing smaller, more efficient generation systems that don't need to cope with massive peak demand. Computers are good at this, providing they can talk to the electrical equipment -- and that means smart, efficient, low speed wireless networking. It's what ZigBee has been built for.

ZigBee is designed to take virtually no power -- the amount each chip will sip from the mains will be tiny even compared to that lost in switches and wiring. The benefits of a ZigBee-enabled network in the home will be much larger: no lights left on accidentally, heating optimised to take power from the grid when it's cheapest, instantaneous reporting and isolating of equipment either faulty or just too greedy for its own good. ZigBee manufacturers are aiming at a cost of around $3 a chip: assuming that the packaging and ancillary costs up that to £3, and even if a ZigBeefied circuit only saves the equivalent of a couple of hours light bulb use a day, it'll pay for itself in under a year.

Of course, ZigBee by itself is no good. It needs to talk both to the computers in the home -- clever power management needs to take an overall view of house consumption -- and run a reasonably smart model. It also needs to communicate with the house owners and be easy to set up: all of which is easiest to do via a PC. For load shaping -- where an electricity supplier provides economic incentives to its customers to run as much off-peak as possible -- the home network needs to be able to negotiate with a central computer. The supplier needs to know who can run what loads at what time, and it then needs to report back with a range of prices so the home systems can plan their most economical usage patterns.

This doesn't necessarily mean you'll need to buy a ZigBee-equipped PC or cable modem. Watch out for the arrival of cognitive radios, chips that can be programmed to do any task from tuning into the BBC to transmitting ultrawideband. One company, Pulselink, has already demonstrated a chipset that can simultaneously cope with UWB, Bluetooth and Wi-Fi -- and the chief executive has been telling anyone who'll listen that it took a handful of engineers six hours to port a completely different UWB standard onto the circuit. The same circuit can also cope with data transmission over cable and powerline -- when it hits production, probably in 2006, it won't be as cheap as a dedicated ZigBee chip but will easily add that capability to the rest without breaking into a sweat.

And because the company is deeply involved in generating standards that let radios describe their standards and operating conditions to each other over the air, a great deal of the practical problems in setting up home networks with differing capabilities - working out who talks to what, and which standards are used - should go away. In theory, not only could a broadband modem with one of these chips spot and automatically communicate with the power company at one end and your immersion heater at the other, but as radio standards improve it will update itself - and any other suitably equipped radios in the area -- with the new capabilities. Another excuse not to take action removed.

There's lots more to be done. Ideas already in use in portable computers -- circuits that can slow down or pause parts of themselves when not required -- should spread to domestic and industrial devices. We should build power supplies that know when to go into ultra-low consumption standby modes, communicating with their host equipment to minimise requirements.

But the bottom line will be wins all the way. We save money, the burden on our overstretched supply systems is lessened, and the IT industry has a whole new and very profitable market to develop. Oh yes, and the chances of hitting the deadlines for emissions before irreversible changes hit the ecosystem are much improved. Who could resist?