- Is the Cisco MARS mission going to abort?
- First iPhone worm spreads Rick Astley wallpaper
- 10 stunning 3D buildings made with Google SketchUp
- Open source software ready for big business
- Four reasons to buy (and one reason to avoid) the Droid
Network equipment vendors and industry watchers are sounding the alarm that RFID threatens to overwhelm enterprise networks with operational demands.
Addressing the issue are big-name companies such as Cisco and start-ups such as Reva Systems, which later this summer plans to unveil a network appliance designed to provision large-scale RFID networks and integrate them with back-end management and security resources and enterprise applications.
The problem is not the volume of traffic that RFID networks create. Rather, it is the sheer number of tags and tag readers that are anticipated. The current RFID approach can't scale to handle those numbers.
"Without an architecture for RFID, large-scale deployments are not possible," says David Passmore, research director for Burton Group. "At the reader level, in dense deployments you have to worry about RF interference, channel assignments and all that RF stuff. Most people don't have the tools and training to do that on their own."
Cisco's biggest enterprise customers are emphatic about their infrastructure requirements for RFID, says Mohsen Moazami, vice president of Cisco's Internet Business Solutions Group.
"They all say, 'If I'm going to install 10,000 RFID readers on my network, you have to ensure they are good citizens on the network,'" he says.
RFID pilots typically involve some tens of readers, installed in a few sites, scanning tags on a limited number of items. The readers radiate a signal, usually in the 900-MHz band, which activates a tag, causing it to reflect some of that received energy back, along with the unique ID number embedded in the tag's tiny processor. That number is passed back to a local server, running RFID middleware and applications to aggregate and manage the data.
Nearly all of these deployments use proprietary protocols between the tags and the readers, and the readers and the server-based software. Equipment supporting EPCglobal's Generation 2 air interface protocol, and the IETF's proposed Simple Lightweight RFID Reader Protocol, is just now being certified.
Dave Husak, Reva's CTO, has a nightmare vision of the next phase of RFID. Imagine, he says, a Fortune 100 applications architect or CIO walking up to the network IT manager's desk and saying "I've just bought 500,000 RFID readers. I'd like them installed and operating."
"I can guarantee you that guy [the network IT manager] has not been at the table during the RFID discussions," Husak says.
|
|||||||||||||||||
What's needed, he says, is a layered architecture for RFID, embodied in an appliance-like controller that dovetails with the current enterprise network. Looking downstream to the readers, the controller coordinates the activities, monitors and manages the radio frequency environment, authenticates readers, and consolidates RFID data. Looking upstream, toward the enterprise network, the controller interfaces with services such as DHCP, presents data to higher level applications and databases, and links with enterprise security and management capabilities.
Comment