RFID: Changing the way we work with data

Radio frequency identification (RFID) is the new generation of automatic identification and data capture. It first appeared in tracking and access applications during the 1980s. These wireless systems allow for non-contact reading and are effective in manufacturing and other hostile environments where barcode labels could not survive.

RFID has established itself in a wide range of markets because of its ability to track moving objects, and is now on a true growth path that will lead to wide adoption through any number of industries.

There are really two stories in RFID. One has to do with the RFID industry itself - the tags, transceivers, antennas, etc. This industry will follow a very predictable format. It will go from early leaders, to major players, to eventual commoditisation. In fact, the industry is depending on this based on their predictions of bringing the costs down to broaden the market applicability. You are already seeing this in major electronic companies and barcode technology companies that are getting into the market, including Texas Instruments and Intermec.

Just as barcode went from expensive exotic to commonplace, the direct RFID industry will do the same.

An explosion of data

But the second story in RFID is really the exciting one. This is a story that impacts everyone in the computing industry. Those who recognise it and know how to leverage it will be in for very interesting times.

The real story behind RFID is the explosion of data that these systems will create. It is easy to see a 10-times or even 100-times increase in the amount of data that a business interacts with. This will lead to some interesting challenges and opportunities in several areas.

With data, there are always four things that you need to do: collect it; move it; store it; and act on it. You collect it through devices, device readers and collection programs. You move it around networks both internally and externally such that the right people and applications in the right places can properly utilise it. You store it for later actions, for audit and regulatory purposes, and for analysis and correlation purposes.

But, of course, none of this matters unless you have the applications and systems that let you act on the data. Fortunately, there are products and technologies designed to assist in all four of these areas.

RFID adds dimension to data

RFID isn't just about more data. RFID will fundamentally change the way we work with and manipulate data.

Typically, systems and applications are concerned primarily with the content of data messages. But with RFID, these systems and applications will have to become more concerned with the additional dimensions of time, place and sequence.

RFID messages are all about what happened where and when. So our transport, storage and application systems will have to evolve to handle these new added dimensions. Here's a quick example: Today, an application might know, in theory, how many of a given item are sitting on a warehouse shelf. But with RFID, that application can receive a message from either the product or the shelf itself, indicating exactly what product was placed on what shelf, when it was placed on the shelf, and when it left the shelf. That fundamentally changes both the message system itself (real-time becomes much more important) and the application that is concerned with the information.

Every business application is, in essence, a set of compromises. Those compromises are between accuracy and data requirements. Every application could be more accurate, and consequently more useful, if it had more data. But too often, the cost of acquiring that data was too high. Finding out exactly what was on each shelf in a warehouse and keeping that data accurate to the minute is simply not practical if every action requires someone to key in the changes at a workstation.

But with RFID, the cost of data acquisition drops dramatically, opening up the possibility of much better applications. Once in my past I was responsible for an application that calculated safety stock levels - how much product should be on a shelf to ensure that we never ran out of an item. The calculation for this was pretty much a rough estimate. It could have been better, but the amount of data required to make more accurate estimates was simply too expensive to obtain - it wasn't worth the acquisition cost. But with RFID messages, the application could know what trucks were on the road, what product was on those trucks, when they could be expected at the dock, and so on. The calculations could have been much more accurate at very low cost.

This dynamic will fundamentally change our business applications and the technology systems that support them. Systems will be expected to sift through millions of pieces of data to produce up-to-the-minute calculations and information related to people, products, locations and business process flows. Message movement systems will be expected to co-ordinate millions of small but important messages, routing them to appropriate applications, storage mechanisms and analysis engines.

Advanced correlation engines will be needed to take raw messages - with information about timing, location and sequence - and make intelligent decisions about their meaning to other processes and applications.

And finally, applications themselves will be expected to make more intelligent decisions with faster reactions to incoming data points. Overnight batch jobs will not be all that useful in the world of RFID.

RFID in the real world

To bring this to the real world, let me relate a recent experience I had while shopping at a major home furnishing store. I was interested in a couple of barstools that were displayed on the floor. The sales rep looked them up on the computer, which said there were three in stock. So what does he do? He goes back into the warehouse to physically look and see if there are, indeed, three in stock. I chuckled, and I'm sure he didn't understand why. But I was thinking of the IT department that had spent, no doubt, millions of dollars building an advanced inventory and logistics system for this guy, just so that he could waste time double-checking the computer.

Now picture the same thing with RFID. Every barstool is recorded as it hits the receiving dock, then again as it is placed on the shelf. If it is moved from one location to another, the RFID system updates the application again. And finally, when it moves from the bin to the floor, or the shipping dock, or through the cash register checkout line, the RFID updates the system again. Finally, maybe, that sales rep can get on to his real job, and not serve as a manual inventory validation system.

So, over time, RFID will change most aspects of how we deal with data. We'll be dealing with much more data, new dimensions of that data and new methods of acting on that data. Companies such as Progress are well situated to take advantage of all this change with systems for the collection, movement, storage, distribution and application of all the information that results from that data.