RFID: Setting the standards

When the Massachusetts Institute of Technology ended its four-year collaboration last fall with dozens of blue-chip companies researching radio frequency identification technology, the group transferred much of its expertise to an emerging standards body known as EPCglobal.

Working with the results -- and many of the same people from MIT's Auto-ID Centre -- EPCglobal is now pushing to create RFID standards using existing communications network infrastructure and the electronic product code, a system for assigning unique identification numbers to items.

Tom Laffey, vice president at integration software maker Tibco and chairman of EPCglobal's Software Action Group, has been with the standards effort since it was at MIT and has seen the hype surrounding RFID reach new levels over the past year. Laffey recently spoke with CNET News.com regarding his views on EPCglobal and the technology's rate of progress.

Q: The one criticism I've heard regarding the EPCglobal effort is that it has yet to publish its standard for the tech specifications to be used in RFID tags. Do you feel this is a valid complaint, and will we see a tag standard published soon?
A: It's a valid concern and, of course, it's both a software and hardware issue. I was at some of the meetings on this recently, and they're about 99 percent of the way there on at least a 96-bit tag standard. That needs to be settled because a lot of other standards will be driven by it. One of the delays has been in making sure that the tag standard is going to work outside the retail, consumer product goods area, because airlines, the Department of Defense and a lot of other kinds of companies outside of (consumer product goods) will want to use it.

The group has basically been in agreement on the standard guidelines for tags for five months. But as things were turned over from Auto-ID Centre to EPCglobal, it became a little more painful. From my point of view, it's done and there's nothing that should keep the people who want to make tags from starting production, as I don't expect any major changes.

Can you talk about why the transition from Auto-ID Centre made it harder to establish the standard?
Part of it is just the logistics in the move. There was a long delay related to how we might treat intellectual property issues. A number of the companies involved had concerns about that. After we resolved those issues, things picked back up.

Work stopped in November and got going again in the middle of February, and it's still a little slow because not all of the companies have moved over from Auto-ID completely, because they wanted their lawyers to have time to review the new intellectual property policies. But I don't think we lost anyone officially. There are some technology companies like Microsoft, IBM and Oracle involved. And if they've gotten past the intellectual property issues, as they have, the problem is pretty much behind us.

What is the next standard you think EPCglobal will undertake?
I've been working more on the software side, but there's been a lot of heated discussion around reader standards, and that's an important one. People need to get to a place where they can interchange their tags and readers, because that's going to help drive down the costs and simplify quite a few things, including integration.

If you look at the PC market and how the Wintel architecture affected the whole ecosystem around the PC, how it drove down costs and increased adoption, that's the sort of direction most people need to move forward. I think the tag and reader standards are equivalent. There will be a lot of important software standards, but you have to have the basics in place. Once we can get the protocols down on how communication between the tags and readers works, I think we'll have achieved something substantial.