Many data centres are increasing asset density to support growing demand for computational power, driven by AI and Internet of Things applications. How are data centres managing all these physical assets?
Higher data centre asset densities
More apps, more connectivity. Billions of people talking to each other via technology, joined by trillions of things. All that interaction needs computation. Add to that artificial intelligence algorithms that need vast and growing computational power.
These trends have clear consequences for data centre infrastructure. They require more cables. More servers. More ports. Higher-density racks. And larger server rooms in bigger data centres.
“How do you maintain efficiency for moves, adds and changes across so many assets? That is a question we regularly hear when talking to data centre managers,” said Linda Chochoy, Product Manager, Market Development, at Brady Corporation, a supplier of automated identification and data capture solutions.
“How do you keep physical asset audits feasible? How do you keep the data centre’s real-world, tangible assets aligned with its DCIM or core asset databases?”
One tried-and-tested way to manage overwhelming and growing numbers of assets is automation using machine-readable serial numbers or barcodes. When scanned, they instantly generate digital data and can connect a specific physical asset to the correct digital record.
The barcode’s more advanced siblings, the Data Matrix and QR code, are even better suited to very high-density asset environments. They allow data parsing and enable multiple automated, fast applications based on the same code. If that is not fast enough, label-integrated UHF RFID technology can provide asset identification at hyperspeed.
Tiny QR or Data Matrix codes
The advantage of barcodes is that, when scanned, they instantly generate a digital data trail. However, barcodes have been around since the 1960s. In data centres, they can take up too much space, require labels that are too large and, as a result, can hardly be used to identify each server port separately.
Due to their size, they also need considerable cable flag labels to maintain code legibility. All those large flags can create a mess, particularly in high-density racking.
QR codes and industrial Data Matrix codes are smaller and able to include more information. They can be printed in millimetres while maintaining scanner readability, allowing them to fit on tiny cable flags and individual ports in high-density servers, routers and switches.
Classic retail scanners are not sufficient to read these codes. They require high-accuracy image-capture devices that can include near-field and far-field technology, or a combination of both.
Advanced Linux-based scanners are available with native JavaScript, enabling hundreds of custom, user-defined applications. To create labels for these applications using tiny, high-resolution Data Matrix and QR codes, high-accuracy and high-precision label printers are needed that can print at 203 or even 600 dots per inch (DPI).
Printers can be set up to print asset codes directly from existing systems as QR or Data Matrix codes on reliable, machine-readable labels.
The importance of tag and label reliability
Some data centre managers have become wary of using labels because of negative experiences. Labels take time to create and apply, and just when they are about to pay off, lower-quality labels can be found at the bottom of the server rack.
“Labels that cannot stay attached serve no purpose at all,” added Chochoy. “A data centre is a challenging environment that requires professional, high-quality labels.”
Server-generated heat can be detrimental to lesser label adhesives, while curved cable surfaces require specialised label materials.
Brady has researched and developed a range of professional labels made from nylon cloth and advanced polymers. According to the company, these labels can keep data centre cables and components identified.
In fact, laboratory testing using internationally accepted FINAT and ASTM methods indicates that the labels stay attached and remain legible for up to 10 years in data centre environments. That is longer than the lifecycle of most data centre assets.
Five outcomes for well-identified data centres
A well-identified data centre has every cable identified, including all fibre, UTP, power and other utility cables. All server, router and switch ports will also have a unique, coded identity.
This enables DCIM or system updates with just two one-second code scans: scan the cable, scan the port.
Well-identified data centres can provide five considerable benefits:
- More accurate and faster moves, adds and changes (MACs), through faster identification of the correct assets, with a greatly reduced risk of unplugging the wrong cable
- Faster troubleshooting, with clear identification on every cable and asset
- Faster rack and server-room audits by scanning physical assets against digital infrastructure records or DCIM data
- Faster, more feasible data centre inventory checks
- Smoother server-rack commissioning and decommissioning
In addition to cable and port identification, many higher-tier data centres also label transformers, pipes and other infrastructure. This can increase the speed and precision of a wider range of interventions that may be required to protect uptime and maintain high service levels.
RFID: a physical asset radar
To achieve even higher asset identification speeds, RFID technology provides another option. Instead of scanning each asset’s code, RFID readers can detect all assets in range in a single scan.
Passive UHF RFID labels do not require batteries and feature thin, integrated antennas that provide a read range of up to 15 metres in open spaces.
Brady offers specialised on-metal RFID labels designed to circumvent known radio-signal distortion from metal surfaces. According to the company, the read range in data centres is approximately 3–4 metres, which is sufficient to confirm the presence of labelled assets across multiple server racks in a matter of seconds.
“The technology works like radar, in the sense that the labels bounce back radio signals emitted by UHF RFID readers,” said Chochoy.
A portable RFID reader enables users to select a detected asset’s ID number on screen and write to that tag or its record. Through signal-strength variation detection, it is also possible to home in on a selected asset using onboard visuals and sounds.
If fixed RFID readers are placed at server-room entrances, every labelled asset moving in or out can be detected automatically, along with its direction of movement.
This makes it possible for data centres to automatically update server-room inventory lists, sound a buzzer or send a text message when the wrong asset moves in or out, and complete server-room inventory checks and audits in a minimum amount of time.
Webinar: Data centre lifecycle identification
How can efficiency goals be optimally supported across the data centre life cycle? Discover the role of reliable identification in data centre construction, commissioning, handover and lifecycle operations.
Topics:
- Design and build: Support compliance, high durability and smart traceability
- Commissioning: Enable high-speed execution
- Handover: Quickly reconcile physical asset labels with digital records
- Lifecycle operations: Protect uptime and facility ROI
Fill out the form to receive a free recording of the data centre identification and traceability webinar.
