Farpoint Group –
As I was writing last week's column on the relay-point model of wireless networking, it occurred to me that I've never written a column on ITworld.com on ultrawideband, or UWB. This is a serious oversight, because UWB is one of the most important wireless technologies ever developed. In fact, I've often said that if radio didn't exist and a bunch of scientists and engineers got together this afternoon to invent it, they would probably start with UWB as a first effort, and one that might carry the field for quite some time.
I'm cheating here, of course, because the first radios were, in fact, ultrawideband in nature. Early radios used a spark gap technique, whereby a very high voltage would jump from one electrode to another nearby, creating a spark and corresponding levels of radio-frequency (RF) energy (the Marconi Calling site has the details). By adjusting the size of the electrodes, the distance between them, and the amount of voltage, one could transmit pulses great distances -- perfect for Morse Code, which was the basis for the earliest wireless communications (data, not voice, you should note). We can think of spark gap as an early form of UWB because the amount of bandwidth used was very great when compared to the amount of information in the signal.
And that's the basis of UWB -- use a huge amount of bandwidth to send a relatively simple signal. If you're familiar with spread-spectrum communications, the basis of modern cellular networks and wireless LANs alike, then you'll realize that UWB is really just an extreme form of spread-spectrum. FCC rules today allow us to use the spectrum from 3.1 to 10.6 GHz (that's 7.5 GHz of total bandwidth) for UWB, provided we keep our power emissions very low -- about one-ten-thousandth of a milliwatt. For you non-engineers, this might seem like a very small amount of power, and it is. But very low power levels like this (and even much lower) are quite common for both cellular and wireless LAN receivers. The problem with UWB is that we're talking here about transmitter power, and, with this power degrading with the square of distance, what we receive at the other end of a UWB connection is very, very tiny indeed -- essentially, it looks like radio noise -- very weak, and very wideband.
OK, enough with the techie stuff. The bottom line is that UWB signals, being very wide, are quire reliable. But since they're so weak, they won't carry over much distance. The wideband nature of these signals means that they have great capacity to carry information, and thus throughput will be high. It just won't be all that high once the distance extends beyond a few meters.
And that makes UWB perfect for wireless personal-area networks (WPANs). You'll recall that this is province of Bluetooth. But whereas Bluetooth has peak performance of about .75 Mbps over a distance of 10 meters, modern UWB implementations can move 110 to 480 Mbps over the same range! UWB's simplicity, relatively straightforward implementation, and potential for low cost make it an interesting element in emerging limited-range applications. Many are now talking about UWB as an interconnect for home entertainment applications -- it could, for instance, connect a flat-panel display or TV to a set-top box (which still won't be capable of sitting on top of any know TV set, but that's a topic for another day). But this is only the beginning -- we can see great potential for toys and games, digital camera links, and other personal-area applications including the relay-point concept I discussed in my last column. UWB promises to be a great success, and may be the third radio in handsets also equipped with Wi-Fi and a wide-area technology like CDMA or GSM.
But all is not completely well in UWB land. A bitter battle over standards has been flaming away for some time. The IEEE 802.15.3a Working Group is trying to reach a physical-layer (PHY) UWB standard, but has been unable to get the votes necessary to finish this task. On one side we have the Multi-Band OFDM Alliance, or MBOA; on the other we have, well, Motorola, which established a presence in this space via the acquisition of UWB pioneer XtremeSpectrum. The battles rages on, and we may have a resolution this month. However, even if we don't, the market will decide what's best. It's not at all clear that we must have one single PHY for 802.15.3a -- there are, for example, eight different PHYs in 802.11 today, with another (802.11n) under development now. Multiple PHYs really don't slow market development, and even standards-based products differ from one another in the quest for customers (we call this the "standards Plus" approach).
I think UWB has a very bright future nonetheless. You'll see a lot of products incorporating UWB over the next few years. It's not, like Bluetooth claimed to be, a wireless LAN. For many personal-area applications, though, it may be just what the doctor ordered.