Its backers claim it will revolutionize supply chain management, transforming everything from transport to document tracking to payment systems.
But some experts warn there’s little margin for error in installing readers, and vendors need to develop better tools for integrating radio-frequency
identification (RFID) data with back-end business software
RFID uses electronic “”smart tags”” (tiny microchips, each with an attached antenna) and readers to assemble and transmit data. When prompted by a reader, the tag broadcasts the information stored as an electronic product code (EPC) on its chip.
Many companies hope to use RFID to reduce distribution costs by tracking billions of objects that move on ships, trucks and planes, and through warehouses and shop floors.
However, some experts caution the operational complexities and challenges of deploying RFID networks need to be addressed.
The high cost of RFID tags is just one of them. Some years ago, smart-label tags were priced between $US1 and $2 each. Today they cost 30 to 40 cents. Analysts predict as volumes increase, the price per tag may fall below five cents.
“”While tag prices seem a major barrier now, they will likely become a minor issue down the track,”” said Predrag Jakovljevic, a research director at Montreal-based Technology Evaluation Centers Inc.
Jakovljevic added problems with RFID implementations can range from “”collision”” of tags to interference from metal racks, liquid items and forklifts.
“”All this requires careful positioning and fail-safe testing of tags and readers alike, especially while read-ranges of high-frequency tags remain quite short.””
Some say the biggest challenge is creating an integrated RFID infrastructure.
For instance, Jakovljevic says, enterprise application integration vendors have yet to develop an adequate range of tools to get data from RFID tags and readers into existing business systems.
He acknowledged the significant strides made by supply chain software vendors such as SAP, Provia, RedPrarie, and HighJump in developing RFID interfaces to their applications. However, he said, even these applications require retailers to deploy specialized middleware and hardware to manage the deluge of data coming from readers. The upshot: companies implementing RFID are often forced to do expensive custom integration work.
RFID network infrastructure vendor Blue Vector Systems says it has antidote to this problem.
The Mountain View, Calif.-based startup has devised a strategy to reduce the labour, complexity, and cost of RFID deployments.
“”It involves the creation of an RFID infrastructure that mimics IP networking paradigms,”” said Anurag Mendhekar, Blue Vector’s CEO.
He said IP networks today are created from homogenous building blocks — routers, switches, hubs and VPN devices — that function regardless of business process or application.
“”You wire these devices together, open up a Web page, set IP addresses for each appliance, hit an update button, and your IP network is up and running,”” he said. “”We’ve extended the same model into the RFID infrastructure realm.””
To this end, Blue Vector has launched its X-3000 product line — a suite of distributed computing appliances that includes the company’s RFID routers, RFID Network Manager Appliances (NMAs), and a Web-based configuration interface.
“”You wire up the RFID routers, connect the NMAs, do a ‘fit to’ configuration, hit the update button…and you’re all set. Then as you start expanding your network, you just add more routers and network managers.””
He said the X-3000 system reports on the RFID network status in two ways. “”There’s a network view that tells you if there’s a problem with a particular reader, for instance, and an intelligent object view that lets you know if there’s something amiss with RFID-enabled intelligent objects — such as specific dock doors in a particular distribution centre.””
The entire system, he said, can be managed from a single customizable screen.
The system’s adaptability, Mendhekar said, is a major selling point.
He added a flexible infrastructure makes it easy for companies to drop in new types of behaviours, models and objects and sweep them across the network.
