As I've said before, RFID is going to change the way companies do business. Combined with sensor networks, it will give unprecedented visibility into the supply chain and will someday give companies the ability to make decisions while goods are in transit -- decisions that could swing millions of dollars to the plus column. There are still obstacles that must be overcome, however; and as they say, forewarned is forearmed.
To get the technical details right, I spoke with Bill Colleran, CEO of Impinj, a fab-less semiconductor company that sells chips to RFID tag manufacturers. He tells me that the obstacles to widespread RFID adoption, none of which are insurmountable, are both political and technical.
If you're piloting an RFID network in your warehouse in a controlled test environment, everything will probably work fine. But in the real world, when you have, for example, 75 cases of beer on a pallet rolling by a reader at 10 mph, problems arise.
At 10 mph, your reader is going to pick up the cases of beer on the outside of the pallet. But what about cases on the inside? There's a good chance they won't be read. The reader's range is too short, the forklift is going too fast, and the cases in the middle are buried behind too many other cases of beer for their tags to be accurately queried.
Part of the problem lies in the tendency of liquids to absorb electromagnetic energy, such as that of RFID signals. And if an RFID signal has a hard time passing through beer, you should see the problem it has with liquid soap. Anyone with such a viscous product is going to have difficulties.
Is there a solution? Yes -- what's needed is not a high-power reader, but a low-power tag. Tags have no power of their own; they come to life when interrogated by a reader and reflect back that reader's wake-up signal. A low-power tag requires less energy to wake it up.
Another obstacle is the "dense reader" problem. Test one reader on a pallet or two of tagged cases, and you'll have no problem. But install short-range readers in a warehouse by every dock door -- say 100 readers -- and the RF energy transmitted by one reader starts to interfere with the others.
The new standard for tags and readers, Generation 2, has modes where it anticipates dense reader deployments, but it has never been tested in volume, Colleran says.
A third obstacle is interoperability. Unlike the Wi-Fi Alliance's work on 802.11, at present there is no interoperability testing going on for RFID. So, a dozen manufacturers can claim compliance with whatever the current standard is -- be it Class 0, Class 0+, Class 1 or Class 2 (Generation 2) -- but tags and readers from different manufacturers may not work together.
The world is heading to the Generation 2 standard, with products expected as early as the second quarter of 2005. If you're contemplating getting a jump on RFID by deploying a Class 0 product now, some vendors offer devices that claim they are upgradable. But Colleran compares that to having a computer from 1995 and deciding to upgrade to Windows XP. You know darn well you're better off buying a new computer than trying to make XP work on an old system. Similar performance issues face the upgrading of older readers.
Ultimately, I can't tell you whose tag or reader to buy. Just remember that, although they are gaining traction, as it stands today these are still new technologies. Buyer beware.