Vendors hope their latest RFID wares will automate the supply chain, but cost, reliability and security issues must first be addressed.
The industry buzz says that RFID is the next great ship sailing to the promised land of automation – in this case, supply chain automation. Last year’s ratification of EPCglobal’s Generation 2 standard (for multi-vendor interoperability between chips and readers) joins the ongoing mandates from Wal-Mart, the Department of Defense and the FDA. With Cisco’s announcement earlier this month of its new RFID add-on board , the technology appears to be on course for mass adoption.
Despite technological progress, serious issues remain, including an almost prohibitive cost (at least for many of the most promising applications), system design unknowns and concerns over reliability and security. Developers are optimistic about RFID’s future but remain cautious about its current capabilities.
“If you look at RFID’s history, you’ll see tremendous amounts of promise and little achievement of those promises … the technology is still immature,” says Daniel Engels, director of research for the Massachusetts Institute of Technology’s Auto-ID Laboratory. “We are still very much in the overhyped phase. As I talk to people implementing, they are disillusioned with the reality of this technology and that has in many ways slowed the adoption and enthusiasm.”
By now most network executives have heard enough of the RFID buzz to understand its ultimate promise. To recap: With RFID, objects are equipped with tiny radio frequency tags that constantly broadcast their identity and whereabouts to strategically stationed readers. With EPCglobal’s Electronic Product Code standard , every tag is a unique identifier. As an object passes one reader after another – or even if it stays in one location – it broadcasts its identity and location. This lets a company eliminate manual processes, including scanning bar codes and tracking shipments with paper documents. When RFID is adopted throughout the supply chain, a slew of new material management practices will be born. Lost shipments hiding out in the wrong part of the warehouse? A thing of the past. Keeping an extra 10% of raw materials on hand to cover those lost shipments? Unnecessary. Manual inventory verification? Gone. Real-time stock reordering? Absolutely.
But with tags priced at about $1 each, according to early adopters, the cost of implementing must be weighed against the value of the object being tracked. Item-level tracking – the issue that causes privacy watchdogs to foam at the mouth – is currently unrealistic. A company isn’t going to spend $1 to give a unique identifier to items that cost pennies. And until RFID is attached to all those bars of soap and razor blades, the idealistic, self-managing supply chain will have to wait.
That means the happiest of new RFID users are those who are looking to do a better job tracking highly expensive items in mobile situations. Consider Brian Taylor, IT specialist for the King’s Daughters Medical Center (KDMC) in Ashland, Ky. Last month, Taylor rolled out an RFID system for tracking devices used in the cardiac, electrophysiology and vascular catheterization labs of KDMC’s Heart and Vascular Center.
The system, from Mobile Aspects, consists of three RFID reader-equipped cabinets – each sporting an embedded computer – and an SQL-based application with Crystal Reports module. The system also is linked to the hospital’s physical security system. Staff members swipe their identity badges to access supply cabinets. They input patient information and remove the needed gear, all of which is tagged. Items are automatically logged. Plus, the system records who retrieved the item and for which patient. If an item is returned, even to another cabinet, the system records its new whereabouts, Taylor says. In this way, the center always has an accurate inventory and charges patients only for equipment used. The RFID cabinets also track the expiration dates of stocked products.
Because the tracked items cost hundreds to thousands of dollars, such as specialized shunts, internal defibrillator and pacemakers, RFID was a practical choice. “At about $1, the No. 1 problem for us with RFID technology is the cost of tags. We’ve made the strategic decision of going with higher ticket items,” Taylor says. “It wouldn’t make sense if we were trying to track $10 items. . . . Five to 10 years from now, the price will be down to 10 cents a tag, and that will change everything.”
Applications and limitations
Creative vendors, systems integrators and users have begun to find uses for RFID beyond managing inventory. However, the physical limitations of the technology can be overlooked in all this new-application hoopla. The tags have a maximum broadcast range of about 10 meters, with range affected by factors such as humidity in the air, researchers say. Accuracy, even under laboratory conditions, is never perfect, though vendors such as Texas Instruments and Philips are working on increasing range and reliability. For instance, TI, which makes the inlay chips that power the tags, has been working to refine tag readability so that they are more consistently read across different product SKUs, at different speeds of travel, and so on, says Enu Waktola, marketing manager of retail supply for TI.
But no matter how large the radio range grows, the nature of mobility means that objects will sometimes move out of range, Engles adds. “It doesn’t matter if we can reliably do it at 10 meters. We’ll want to do it at 10.1 meters, and then we’ll get upset if we can’t read 100% of tags,” he says.
While making an inherently unreliable system dependable is an age-old infrastructure issue, many untested would-be RFID consultants haven’t got the experience to do so, industry experts say. Even if the physics are managed, other design quandaries remain.
The pharmaceutical industry is a case in point. The FDA is pressuring distributors and manufacturers to implement RFID technology by 2007. Most major pharmaceutical manufacturers and distributors are exploring pilots. Yet, even a pilot can be hard to design.
“Pharmaceutical manufacturers are faced with some complicated infrastructure questions. One of the most paralyzing factors is how do we institute [a drug’s] pedigree? When we start embarking on how we create, transfer and authenticate pedigree, we start talking about massive IT investments and data integration projects all the way down to data construct for the RFID tags themselves,” says Dennis Kim, senior manager of supply chain operations for a U.S. drug maker.
Then there are interference issues. For instance, HP wanted to use RFID to track thousands of servers and blades in its internal data centers as they get shuttled around for upgrades and repairs. But first the scientists at HP Labs had to figure out how to keep the metal racks that house the servers from causing too much interference, says Salil Pradhan, chief technologist for RFID technology at HP Labs. The scientists succeeded, but their solution requires specially enhanced readers. Most enterprises don’t have the luxury of an internal IT R&D department to re-engineer RFID products.
Pradhan says security is another weak spot. After the tags transmit their data, it becomes ordinary data and can be protected by ordinary security measures. However, little can be done today about an attack that intercepts radio transmissions. Adding encryption to the tags would cause them to grow in size, in energy requirements and, unacceptably, in cost.
But the dangers of interception are overrated, says Paul Strzelec, director of marketing for the EPC network services team at VeriSign. He likens tag broadcasts to a car’s license plate. The number isn’t secret – the information the number represents is. And that is safely stored in a database. Protecting it requires nothing more than such application security as access control and authentication.
The real danger, he says, is in how companies will protect their proprietary warehouse information from wrongdoers while still sharing it with suppliers and customers to automate the supply chain. To that end, VeriSign offers managed EPC services for naming objects (similar to its DNS service), and for sharing EPC data.
On the upside, data volume is no longer expected to kill your network, as pundits once feared. “I don’t think RFID is going to overwhelm IT systems anytime soon. Even if every object gets an RFID tag – 100 billion tags, each read 100 times – 100 billion bytes is not that much data. Compared to video or other files, this is noise,” says Pradhan. Moreover, earlier this month Cisco announced an add-on board that will ease integration of RFID traffic onto the data network. (See related story .) Plus, RFID data aggregation is becoming better understood, so with an appropriate design, your storage systems shouldn’t be overloaded either. Tags continuously broadcast information, but that data isn’t noteworthy until the tag changes location. Then “all you need to know is that this box came at this time, or left at that time – a high-school kid can write a piece of software to handle that,” Pradhan quips.
As for the business analytics software that will mine the data you do store, help is available. It comes in the form of middleware from start-ups such as OATSystems (one of Network World’s 2005 Start-ups to Watch ) and stalwarts like Sun’s Sun Java Systems RFID and Microsoft’s RFID Solutions Architecture, among others.
Learn more about this topic
Copyright © 2005 IDG Communications, Inc.