Spectrum Policy

Excerpts from a special report from the Economist:

"There is no limitation in the spectrum of sound waves, only in the refinement of the human ear. The same can be true in the electromagnetic spectrum.

There are four broad categories of new technologies that could make this idea a reality. The first is called “spread spectrum”, or “wideband”. As both names imply, this is a way of spreading an electromagnetic signal across wide bands of frequencies at low power, instead of booming a high-power wave through a narrow band. Wi-Fi is one good example of wideband technology—the large range of frequencies and the low power allow it to co-exist with cordless phones and other devices. Hopes are highest, however, for a new technology called “ultra-wideband”, which will communicate by whispering its signals so softly across the frequency bands of other, higher-power transmitters, such as broadcasters, that these will not even notice the presence of another signal.

Another approach is to use “smart” antennae. These are systems of multiple antennae that can “aim” a signal in a particular direction (instead of radiating it out indiscriminately) or pick out a particular signal from background noise by calculating the wave's angle of arrival (for example, from a satellite instead of a source on the ground).

A third technology is “mesh networking”. In a mesh, each receiver of a signal also re-transmits it. Every meshed laptop computer, for instance, in effect becomes a node or router on its network. This has three advantages. One is that, as with spread spectrum, signals can be sent at very low power, since they only have to travel to the next user's node, which will be hundreds of metres, instead of kilometres, away. Another is that each newcomer to the network not only uses, but also adds, capacity. A third is that the network will be robust, since traffic can be re-routed easily if nodes fail, the approach already taken by the internet.

Open-spectrum enthusiasts are most excited, however, about the day when radios become software-powered computers, or so-called “cognitive radios”. This would end the limitations of dumb radios. “Moore's law meets Marconi's transmitter,” says Kevin Kahn, research boss for communications at Intel, the world's largest semiconductor-maker, referring to the prediction, so far correct, by Gordon Moore, one of Intel's founders, that the number of transistors on a chip doubles every 18 months. Radios would double their intelligence every year and a half, in other words. They could learn to hop around on the spectrum to find quiet bands for transmission, to encode digital information in new wave forms, or to analyse incoming noise and pick out only the relevant signal".