Power Lunch: It's time for Open RFID
The software industry had Linux. What will RFID users have to "open" up their systems? It's time to put enabling technology in the hands of the masses.
By Daniel Deavours, Director, RFID Alliance Lab | 06/27/2005
Radio frequency identification (RFID) is the future -- at least that's what everyone says, and I believe it. But RFID equipment sure seems expensive. Last fall I bought a high-end multi-protocol reader for close to $3,500. It sure does a lot of nifty things, but all I really need it to do is read RFID tags.
While doing research with the Alliance Lab, we've discovered some very poor behavior for tags right at the edge of the read field. We believe that it's an artifact of particular implementation choices.
As a researcher, I want to look at what's going on "under the hood." I have ideas that I think could improve performance considerably. But I can't test those ideas because everything is proprietary.
If you haven't heard, the trends in radio designs are increasingly moving toward something called "software-defined radios." Basically, it means that the RF hardware is as simple as possible, and nearly all the signal processing is done in software. That makes radios a lot more flexible.
The RF module is tied to a particular frequency band, but the rest is software. I haven't cracked open the case, but my guess is that my $3,500 reader is really just a software-defined radio. Everybody knows about open software from the popularity of Linux and other successful open source projects. Open source is a way for people to share ideas and technologies and for ideas to rapidly evolve. Sure, people still pay for Microsoft Windows, Word, and PowerPoint, but Linux has found a niche, from my desktop systems to the Linksys access points in my home to, yes, even my $3,500 RFID reader.
What I propose is starting an "OpenRFID" initiative similar to that of the GNU Radio (www.gnu.org/software/gnuradio
). Basically, OpenRFID will be the software that drives a software-defined radio to issue basic RFID commands, like read a tag ID. EPC RFID specifications are open and have been submitted as an ISO standard. Once we have the basic protocols worked out, the cost of a simple reader should be no more than the cost of hardware.
I haven't figured out how to fund it or the right open software model (there's a dizzying number to choose from, each with pros and cons). But I have figured out that a good RFID reader should cost a lot less than $3,500. In fact, I priced building my own portable, battery-operated prototype in low quantity, using expensive, high-end components and custom boards, for about $2,000. In quantity, I estimate the cost well under $1,000, and I've seen similar radios for less.
The commercial impacts of OpenRFID are clear: cheaper RFID readers, but it goes much deeper than that. Opening up RFID puts technology in the hands of the masses. It means universities can study and explore design tradeoffs, and that means better technology down the line. It means companies can develop new technologies and uses for readers without having to reinvent the reader or pay expensive royalties. That means more innovation and more companies bringing innovative ideas to market. It means more choices in RFID equipment. It especially means products that meet demand, not products that have the highest profit margins.
Sound interesting? Think you'd like to participate? Let me know.
Daniel Deavours is a research assistant professor at the University of Kansas and director of research of the RFID Alliance Lab
). He can be reached by e-mail at firstname.lastname@example.org