Using RFID to boost business’ bottom line

Plans to deploy radio frequency identification technology in supply-chain settings have kept retailers and manufacturers busy over the last couple of years, spurred by adoption mandates from Wal-Mart, Target, the U.S. Department of Defense and others.

Early IT pilots focused on the physics of RFID – finding ways to improve reader performance, figuring out how to apply tags and other device-focused issues. These days, IT executives are turning their attention to higher-level RFID challenges, such as how to make business sense of the data being collected.

To glean usable business intelligence from RFID data requires middleware, experts say. RFID middleware sits between the readers that collect tag data and the business applications that eventually will digest the data. The software manages RFID devices and turns the data they generate into information that business applications can consume.

Early adopters first worked on RFID devices, and now they’re ready to think about software, says Sharyn Leaver, a vice president and research director at Forrester Research. It’s a natural progression, she says. “Without the tags and readers working correctly, there was no need to think about middleware or applications. As the technology matures, we’ll see the interest shift from hardware up to software,” she says.

As RFID users turn their attention to software challenges, a slew of vendors are standing by with products. There are pure-play RFID vendors, such as ConnecTerra and OATSystems, and retail industry specialists such as Manhattan Associates. Big infrastructure software companies IBM, Microsoft, Oracle and Sun also offer RFID middleware, as do integration specialists such as webMethods and Tibco Software.

Most recently, Sybase subsidiary iAnywhere Solutions got into the race. Last week, it announced RFID Anywhere , a software platform that includes connectors and controllers for communicating with RFID hardware, and simulation software for assessing the potential impact of RFID data loads and content on networks and applications.

ProPath is one of the first users of RFID Anywhere. The Dallas pathology services company uses RFID technology to track biopsies and other pathology specimens as they advance through the laboratory’s complex processes. An RFID tag placed on every specimen that comes into the lab will store information about the specimen, and who performed each task and when.

RFID Anywhere helps track specimens more closely than the lab’s current system, says Mel Lively, director of IT at ProPath. ProPath’s current anatomic information system documents four stages of a specimen’s analysis, while RFID Anywhere will let the lab document nine stages.

How it works

At its most basic, RFID middleware handles hardware integration. With connectivity standards still in flux, there’s no set way to build an interface to an RFID reader, and some companies will want to deploy different readers from different manufacturers. “For now, the reader manufacturers don’t all conform to a single standard, so middleware is an important buffer layer,” IDC’s Leaver says.

Middleware also can provide tools for managing RFID devices. IAnywhere’s software, for example, can communicate with devices through Ethernet or serial connections and push firmware upgrades as required, says Chris Foley, director of RFID at iAnywhere. It provides a Web-based console to remotely monitor the status of RFID devices. Alternatively, iAnywhere can feed device information to a third-party systems management platform, Foley says.

Filtering data is a key role of middleware. It processes the streams of data coming in – not just from RFID readers, but potentially from bar-code scanners or other wireless sensors, such as temperature sensors.

Companies have worried that RFID readers that constantly poll tags will take a toll on network performance and flood applications with unnecessary data. Industry experts say those fears are overblown.

“There’s no question the amount of data pumping through the network is going to increase, but the initial horror stories about it bringing down every network because there’s so much data are overstated,” Leaver says. “People are forgetting about the role of middleware.”

Namely, middleware will inspect the data and apply business rules to get rid of extraneous data reads, and pass along to enterprise systems only the important information or the exceptions that could require someone to take action.

RFID middleware also assumes the role of electronic product code (EPC) commissioning. An EPC number is a serialized product code that uniquely identifies the product being tagged. Users can tap middleware to connect computers that look up EPC numbers to a central directory governed by the nonprofit EPCglobal and managed today by VeriSign – much like the DNS directory VeriSign manages.

Choosing a vendor

When choosing a vendor, don’t think about RFID middleware as a product selection problem, think about it as an architectural design issue, Leaver says.

Companies should plan to deploy an RFID architecture with multiple software tiers to handle different filtering and routing requirements.

“It makes sense to do some filtering out close to the reader, so you’re not sending a bunch of data across the network for no reason,” Leaver says. That being said, you don’t want to have all the business context and business logic of a corporation out at those endpoints. Once the simplistic filtering is done at the edges, more complex data packaging and routing can occur closer to the enterprise networks, where the middleware can access business logic, she says.

The platforms that companies choose need to be reliable, scalable and intelligent enough to understand business processes and how to interact with business applications. “Getting that combination is a little difficult,” Leaver says.