Still many disparities before all the RFID dots are joined

Many manufacturers and retailers believe radio frequency identification (RFID) technology will help them better track their inventory, allowing companies to move to just-in-time production, which will reduce costs. There’s been a lot of recent hype surrounding RFID, but despite some promising signs that the technology is ready to roll, there are still some significant hurdles RFID will have to overcome before becoming part of the mainstream.

RFID technology creates a new layer of network infrastructure — the RFID layer. This layer consists of the RFID tags (micro-transceivers that transmit a stored serial number and other data), RFID readers (receivers), RFID servers with middleware, and the enterprise data network. At the physical layer, RFID readers send out an RF pulse for tags to respond to. The information from the tags is passed from the readers to the servers via Ethernet connections in the data network.

EPCglobal, a not-for-profit group working on electronic processing codes (EPC) and RFID standards, is creating a universal RFID information storage system for companies to share information on RFID tags via the Internet.

Sounds simple enough. So what’s the problem? The reasons for the delay in widescale adoption of RFID are rooted in some familiar networking problems: incompatible standards, device incompatibility, cost and potentially significant network issues. Each of these areas needs to be addressed before RFID can become a widespread technology.

As with the early deployments of wireless networks, there are no universally accepted standards for RFID tags. The ISO and EPCglobal are each proposing a different set of standards. The ISO is pushing for more functionality, while EPCglobal promotes simplification for lower costs. Both groups have companies with patents that can hinder other companies’ product development. Without universal standards and patent sharing agreements, there is significant investment risk to anyone trying to choose a technology for deployment.

Due to different uses and standards, there is a proliferation of different types (Class 0 to Class 5, and now a GEN 2) of RFID tags and associated costs. Different tag types are suited to different uses for different products, like working over long distances or communicating through obstacles. Costs vary from $0.50 to several dollars per tag, depending on volume and functionality. High costs make tags unattractive for inexpensive individual products, so manufacturers consider tags only for cases of products, complete pallets or expensive products.

The tags described are passive, meaning they respond only when hit with RF energy, sending out a signal. As well, there are more expensive, self-powered active tags (in one case using Wi-Fi). Proliferation of tag types means different readers need to be deployed. A retailer using RFID technology would need overlapping RFID readers for each tag type. Since readers aren’t inexpensive ($1,000 each in some cases) and have limited range (six to 30 feet), building a network could require thousands of readers at significant cost.

The enterprise data network is another issue. RFID readers pass along tag information individually for every tag read. Depending on how many items are tagged, and how often they are read, the ensuing traffic could put a significant load on the data network, similar to online transaction processing systems. If real-time inventory data becomes a critical part of the business process, then the data network needs to be designed to handle RFID tag traffic with appropriate QoS to ensure delivery. In many cases the network that the RFID layer would use is not up to the challenge. Also, the future burden of looking up every RFID tag on the Internet will place a bigger demand on the network.

Given the potential for large returns in RFID investment, many manufacturers, distributors, retailers and end customers would like to see the technology succeed. To that end, there are several bright spots. An RFID vendor consortium was formed in August to pool patents to make development easier. Unfortunately, one of the largest players, Intermec, did not join. Intermec holds 145 patents. For RFID to succeed, everyone needs to agree to allow development to make the technology universal.

Large end users are pushing for more RFID deployments. Walmart has mandated RFID use by its 100 top suppliers. Competitive forces mean other retailers will watch Walmart closely, adopting technology if Walmart succeeds. The U.S. Departments of Defense and Homeland Security are also both advocating RFID. Given the need for governments to share security information, it’s only a matter of time before the same RFID initiatives appear in Canada.

Costs are dropping as volumes have climbed. Recently, a large RFID manufacturer publicized an order for 100 million RFID tags for a Korean company. For prices to continue dropping, volumes need to increase further. The only way to assure this is by standardizing the technology so that the same type of components is used in RFID manufacturing. As long as there are many RFID options, there will also be higher costs.

The impact on the enterprise data network is being addressed in significant ways. Networks are getting smarter, recognizing the type of traffic traveling across them and adjusting for that traffic. Cisco, Enterasys, Extreme and Foundry all have technology that should recognize, protect and deliver RFID data. Resiliency in networks has also improved due to VoIP implementations that ensure delivery. New middleware applications for RFID servers should consolidate delivery of data to back-office applications.

While these improvements mean RFID will continue to improve, it won’t likely be widely deployed at the product level any time soon. For most organizations, a pilot implementation of RFID that is confined within their own organization probably makes the most sense for the next year or two. More of the issues facing RFID will need to be addressed before RFID is deployed widely.

QuickLink: 059315

–Kanellakis worked at Enterasys Networks and its predecessor Cabletron Systems for almost 15 years. These days he is enjoying spending time with his family and looking for interesting ventures to pursue.

Related Download
EMC Data Protection For VMWare-Winning In The Real World Sponsor: EMC
EMC Data Protection For VMWare-Winning In The Real World
Download this white paper for a deep dive analysis based on truly real world comparison of EMC data protection vs. Veritas NetBackup for VMware backup and recovery.
Register Now