RFID (radio frequency identification) plays a unique role in Operations. Friday is a relaxed day at Real User Monitoring blog, a time we try to look behind the news at longer-term realities that bear on Operations: perfect for RFID.
RFID is an umbrella initialism for a number of distinct but related technologies and applications. The basic idea is this: a miniaturized “tag” transmits by way of radio a message on the order of, “My name is X3176A88G …” Some tags embed their own batteries, and have finite (but surprisingly long) lifetimes; some are able to operate “passively”, recycling radio energy they receive, and thus communicate essentially eternally. Tags operate on a number of different radio frequencies, to meet specific engineering requirements. Some tags can report not only their identities, but a bit of history or state, such as time-since-last-maintenance or thermal-stress-detected.
Tag them all
Put a tag on every piece of hardware that comes into a datacenter or Operations node, and:
- theft prevention is easier;
- inventory management becomes far less time-consuming; and
- building security tightens up.
RFID provides unique value in these domains. No other technology binds physical assets with digital representations as inexpensively and reliably as RFID.
RFID also can supplement or replace other techniques. A recent article on “greening the datacenter” cited a savings of “more than $40,000 a year” at the datacenter of the Franchise Tax Board in Sacramento from temperature regulation. Strictly speaking, such savings don’t require RFID; well-made digital equipment has been reporting temperature measurements for years by way of SNMP. As a practical matter, though, RFID-transported readings from environmental probes have a couple of advantages:
- RFID-based systems probably better suit practices in Facilities; environmental SNMP messages are too likely to be subordinated to DevOps priorities and styles. Also,
- For historical reasons, it has turned out to be far cheaper to order handfuls of RFID tags to be mounted where needed, than to configure, order, and provision properly-equipped network elements.
Exploitation of RFID in the datacenter was a rising trend about five years ago, as articles in Computerworld, Infoworld, and other magazines attest. What has happened since 2008? Mostly, RFID has “normalized”: it’s become part of the best practices that good Operations staff assume. A rich marketplace of vendors and consultants supplies everything from individual tags to turnkey contracts for making the most of RFID.
Recent years have also seen more awareness of “ghost equipment”: machines that are plugged in, but no one is sure where, and they’re doing no useful service. They draw power, heat your space, and, as just happened in one of our datacenters this week, consume networking bandwidth fighting off crackers trying to get in.
If your facility is not already making the most of RFID, think through a few of the use cases where RFID can save you time and eliminate distractions. Have you ever had to search for a network switch installed during an emergency, left in place, and effectively lost until the next emergency? Do you have a method to track down mislocated, pilfered, or experimental equipment that doesn’t involve the adjective, “exhaustive”?
As with many automations, the biggest benefits RFID holds for the datacenter are likely to be unrelated to RFID directly. Once you have streamlined your practices so that you can take inventory, or pass by a security gate, or correlate the environmental status of several distinct items, in a standardized, replicable way, your freedom from those distractions is likely to liberate your thoughts to attack the next, and perhaps even more important, problem. What will it mean to you to have the physical locations of your assets under digital control?