RFID to Complicate Database Administration

RFID is a burgeoning technology that seems to be right on the brink of enormous success. Basically, RFID (radio frequency identification) is a cheap, chip-based technology for emitting radio frequencies enabling quick, wireless identification of items. Wal-Mart has been in the news recently regarding its implementation of RFID technology to improve its inventory control processes.

By tagging inventory items with RFID tags locating and identifying materials in inventory can be simplified and made more cost-effective. Well, that is the hope anyway. The cost of RFID tags has decreased, with single tags costing about 25 cents. In bulk they can be purchased for as low as 10 cents each, but many industry pundits believe they will have to drop below 1 cent per tag before the technology becomes cost-effective for widespread adoption. That is probably at least a decade away.

But there will be smatterings of RFID implementations here and there. Today, the RFID tags may be applied to pallets of goods, instead of each individual item.

As RFID adoption expands, the information in the tags will need to be stored - and where else should it be stored than in databases? This raises some questions though. We are already storing a lot of data - more and more every year actually. We may store the same piece of data multiple times - oh, it came from a flat file in the production system and we moved it to IMS in the 1970s, then to DB2 in the 1980s, and maybe even to Oracle in the 1990s. And it still "lives" in all three databases. And it gets copied to the data warehouse for analytical processing and those users in product management download the data into spreadsheets every week.

And when RFID technology is adopted maybe we'll have even more data to store. Should a record be kept of every time an item with an RFID tag is scanned? Or just a date (or indicator) for the last scan? Should the RFID be the primary key? Or should it be something more recognizable to end users?

What about privacy concerns? What (or who) can be labeled with an RFID tag? Can we RFID tag students in a school so we can locate and track them in case of an emergency? Or is that immoral (or illegal)? And how much data can we keep if we tag people? Will it be possible to track the movements of a person throughout the day? Today's RFID technology is not capable of such tracking because the signal is not very strong - but what if that changes? The European Union recently announced privacy guidelines that address some of the current privacy concerns of RFID adoption.

Security is an issue, too. If we keep such in-depth records of the location of people or inventory, who can access that data? How long can it be kept? What if it gets into the wrong hands?

But what should DBAs do about RFID today? Well, you can prepare for the inevitable. Read up on the technology so that you understand its capabilities and limitations sp that when it becomes an issue at your company you will be prepared to immediately assist. Keep your eyes open for news articles and web postings talking about the subject. When you hear about companies in your area that are implementing RFID technology, take the DBA for that company out for some liquid sustenance and discuss the project with him (beer) or her (daiquiri). Sometimes the cost of knowledge is only a few drinks!

Before signing off, I want to make a prediction here. Soon, you'll start hearing about RFID databases (like we heard about object databases when OO was hot, and XML databases now that XML is hot). Beware! These DBMS products are unlikely to be based on any sound underlying data model. Instead, it is a probably a quick attempt to tag along onto a hot technology and pick up some sales. I personally see no reason to build an RFID DBMS - our existing DBMS products should be up to the task of capturing and storing RFID data (although some future extenstions to existing DBMS products to support RFID would be fine).

So, DBAs, suck up everything you can about RFID because it will eventually be on your "to do" list.

I just received a great book that can be used to help you learn about RFID technology. It is titled RFID Sourcebook by by Sandip Lahiri. Just like the book cover says, this is a "realistic, no-hype guide to RFID evaluation, planning, and deployment." Check it out if you are at all interested in RFID and its potential.