RFID middleware helps you use captured data effectively: mountains of data won't help you if your information system can't process it efficiently

Food & Drug Packaging, June, 2005 by Graham Wiemer

Most mid- to large-size companies have decided to jump on the radio frequency identification (RFID) bandwagon, seeing it as an inevitable leap in technology that will eventually make their businesses more profitable--even if the initial cost of an RFID system won't show a return on investment for quite some time.

In some cases, these companies didn't actually jump so much as get pushed by Wal-Mart, Target, the Department of Defense and other enterprises that declared RFID tagging would be mandatory for those wishing to do business with them. Regardless of the reasons for using an RFID system, every company needs to find a way to take the flow of information it receives from RFID tags and make sense of it.

That's where RFID middleware comes into play. The amount of RFID information and the speed at which it can be gathered can quickly overwhelm a company's current information system or warehouse management system (WMS)--not to mention the people trying to make sense of it. Middleware is designed to cut through the confusion and make all the data a company receives through its RFID readers more meaningful.

The information from tags goes through a reader to either a Site Manager or Edgeware, which filters the data and routes it to the next layer of software, which should be a middleware platform. "The raw data floods would overwhelm existing Enterprise Resource Planning systems as they're designed today," says Mark Nelson, director of corporate communications for Savi Technology Inc., which holds 13 fundamental patents on RFID technology. "Users would be confused by too much data. It has to be refined and made intelligible to the user beforehand."

RFID tags can be encoded with any information a company desires, but most companies are adopting the Electronic Product Code (EPC), developed by the EPCglobal network. The advantage of using the EPCglobal data structure is that--like the Universal Product Code (UPC) data structure--it will be a common format used across retailers and suppliers, so everyone will know how to properly read and interpret the data inside the tag because everyone will be a subscriber of the EPCglobal database.

Tony Sabetti, director of retail supply chain products for Texas Instruments' RFID division, says the big challenge that companies face is to take this identification number that pops up on an RFID reader and determine the best place to route that information.

"A lot of what the middleware folks are focused on now is building applications that are very efficient at moving the EPC ID numbers from one place to another," says Sabetti. "If you were to take an RFID reader and have it just continually looking for tags and then sending that information off to a master database, you'd pretty quickly overrun that database."

The goal of middleware is to filter, sort and report events that are meaningful.

Sabetti explains, "An RFID reader is constantly looking for information, so the trick in this software is just to read an item once. RFID readers can read hundreds of items a second, and maybe read the same tag seven, eight times within a period of a few seconds. Deciphering that and sending that to a network so that it's meaningful information is the key."

Middleware tries to prevent the RFID reader from completely swamping your warehouse management system software, because WMS software is basically designed with the expectation that one event will be reported at a time. So middleware takes the RFID information and puts it into a form that the WMS can handle and people can put to the best use.

Companies need to make a decision at both the tag level and the data coding level about how they're going to use the tagging information. "We have a lot of customers that are using tag information to reduce shrinkage in their inventory systems," says Sabetti. "For that to be effective, they do need to read the serial numbers, because they need to be able to know when a case of product goes missing. So, as they move product from one location to another, on or off one truck, they need to make sure they account for all the serial numbers."

RFID tagging can be tailored to meet the specific needs of different industries. "We have companies that are thinking about using RFID tags to reduce counterfeiting, to prove that a product is authentic and not a knock-off," Sabetti says. "That requires a security structure on the tags that's much more sophisticated than the EPC Generation 2 data structure, which is not really robust against copying--it's not really intended for that. We're expecting this to be in the neighborhood of 2K-bits (2,000 bits), and we're expecting this to just be used in anti-conterfeiting."

Middleware becomes even more essential when data-rich active RHD tags are used. These tags are best suited to track assets or shipments throughout the end-to-end supply chain.

"The supply chain is so complex and involves so many interactive partners," says Nelson. "Only active tags can store enough information to be meaningful for recording and updating changes along a shipment's journey, such as the route status, custodial responsibility or even condition of shipments, if sensors are used to gauge such factors as temperature, humidity and light."