RFID learns a trick from Ethernet

In writing about NCR’s new branch bank RFID portal (see “NCR takes RFID toward the creepy side”) a couple of weeks ago, I wondered how “an archway equipped with RFID readers that scan information from RFID microchips embedded in a bank card, cell phone or PDA” (from NCR’s press release) could differentiate among the myriad of cards in your wallet or on your keychain. Last week I spoke to NCR’s CTO for Financial Solutions Mark Grossi, who pointed out the possibility of a problem when two or more customers walked through the portal at the same time.

Discriminating among and between cards from different issuers (i.e., accepting “your” card and rejecting “some other” card) can be handled through the use of different wavelengths and discrete identifiers. But when there are two or more cards from the same issuer carrying the same identifier broadcasting on the same wave length – what do you do then?

Now admittedly I haven’t followed all of the technological underpinnings of RFID technology but when Grossi said that the “collision avoidance” was used to prevent that scenario I immediately asked – “collision avoidance or collision detection?” Turns out we both had done stints in the Ethernet card business, back when it really was Ethernet as Bob Metcalfe had invented it and it was really called CSMA/CD (Carrier Sense Multiple Access with Collision Detection).

In a nutshell, the “CD” involved the Ethernet card listening on the wire for a break in the traffic and then inserting its packet. If two stations attempt to transmit simultaneously, this causes a collision, which is detected by all participating stations. After a random time interval, the stations that collided attempt to transmit again. If another collision occurs, the time intervals from which the random waiting time is selected are increased step by step. Today’s Ethernet uses different methods of communication in which collisions are a thing of the past. But the knowledge gained from the CSMA/CD (and the lesser used CSMA/CA – CSMA with Collision Avoidance) appears to be getting put to good use with RFID.

Different RFID manufacturers may use different collision avoidance technologies, but one fairly common one is defined under the name “Tag collision” – “…when the reader recognizes that tag collision has taken place, it sends a special signal (a “gap pulse”). Upon receiving this signal, each tag consults a random number counter to determine the interval to wait before sending its data. Since each tag gets a unique number interval, the tags send their data at different times.”

So just like the old Ethernet networks, the RFID “network” can handle what might be thought to be confusing signals through an adaptive technology that’s easy to implement, easy to explain, yet difficult to improve upon economically. That’s a neat trick.