Asset tagging is fixing a durable, scannable label with a permanent identifier to each piece of IT equipment so the physical device always matches its record, the foundation of the enterprise asset inventory in CIS Control 1. A good standard covers what gets tagged, label material, placement, the numbering scheme, and when tagging happens.
Tags fail in predictable ways. Adhesive gives up in a hot aisle. Thermal-printed labels fade under years of UV from overhead lighting. A tag placed on a removable component walks off with that component. None of these are exotic failures, and all of them are cheap to prevent at the point the tag is applied.
What IT Equipment Should Get an Asset Tag?
Every chassis should carry a tag, and so should every data-bearing component. Beyond that, the first decision is granularity, and it drives cost for the life of the program.
Tagging at the chassis level is the common baseline: each server, switch, storage frame, and appliance carries one tag. It is enough to answer where equipment is and what it is worth, and it is the minimum that supports a physical inventory.
Component-level tagging extends to drives, memory, GPUs, and line cards. It costs considerably more to maintain and it is the only way to track components that move independently of their host, which matters for the same reason it matters in any asset lifecycle program — a drive that leaves a decommissioned server and enters a different one has changed its risk profile, not just its location.
Data-bearing components deserve tagging regardless of the general policy. When a drive eventually reaches disposal, the certificate, the kind of sanitization record NIST SP 800-88 calls for, needs to name a serial number, and that serial has to have been tracked from the moment the drive entered the estate. Retrofitting it at disposal time is guesswork.
Consumables, cables, and rails generally do not justify individual tags. They justify labeling, which is a different thing with a different purpose.
What Asset Tag Material Survives a Data Center?
Polyester or polyimide stock printed by thermal transfer survives; paper and direct thermal do not. A tag specification is short and worth writing down, because the alternative is whatever the vendor shipped.
Polyester or polyimide face stock with a permanent acrylic adhesive survives data center temperature ranges. Paper labels do not. Where equipment will see solvent cleaning or outdoor staging, the substrate matters more than the printing.
Thermal transfer printing with a resin ribbon resists abrasion and UV; direct thermal fades, sometimes within a year. This single choice accounts for a large share of unreadable tags.
Include both a human-readable identifier and a machine-readable code. A barcode alone is useless when the scanner is upstairs. Code 128 is adequate for linear barcodes; a 2D code such as Data Matrix holds more in less space and scans reliably at the angles a rack forces on you.
Tamper-evident stock is worth the premium on high-value assets. A tag that voids when lifted tells you something a plain label cannot.
Where Should Asset Tags Go on Rack-Mounted Equipment?
Tags belong on a chassis face that stays visible once the unit is racked and cabled. Placement is where good specifications fail in practice. A tag applied to the rear of a rack-mounted server is invisible once the unit is racked and cabled, which means every audit requires pulling equipment.
Put the primary tag on the front bezel or on a chassis face that remains visible in situ. Where the front panel is removable, that is the wrong surface — the bezel and the asset are separable, and they will separate.
A second tag on the rear or top is cheap insurance and speeds decommissioning, when equipment is accessed from behind. Use the same identifier on both. Two tags with different numbers on one device is worse than one tag.
Avoid surfaces that get hot, surfaces that flex, and anywhere a tag obstructs airflow, a vent, or a service label the manufacturer needs. Keep tags off the parts a technician removes routinely.
How Should Asset Tag Numbers Be Structured?
An asset identifier should be opaque, sequential, and permanent. That is a deliberately boring specification, and it is boring for a reason.
Encoding meaning into the identifier — location codes, device types, department prefixes — creates a number that becomes wrong the moment anything changes. A server tagged with a site code and then relocated now carries a misleading identifier for the rest of its life, and staff will trust it because it looks authoritative. Location and type belong in the record, where they can be edited.
The identifier should never be reused. When an asset is retired, its number retires with it. Reuse destroys the historical record and produces audits that cannot distinguish the device disposed of in 2023 from the one in production today.
Serial numbers are not a substitute. They are assigned by manufacturers, they collide across vendors, and they are frequently unreadable on installed equipment. Record them in the database, and tag with your own number.
When Should IT Equipment Be Tagged?
Tag equipment on receipt, at the dock, not on deployment. The point where an asset enters the organization is the only point where tagging is reliably complete. Equipment tagged at deployment leaves a gap — anything received and not yet deployed is untracked, which is exactly the population most likely to go missing.
Build it into receiving: the asset is tagged, the record is created, and the serial is captured before the box leaves the dock. This is also when the purchase order, cost, and warranty dates are easiest to attach.
During a relocation, tagging discipline is what allows a physical inventory to reconcile against the record at both ends of the move. A crate that arrives with three untagged devices is a crate that generates a week of investigation, and moves are precisely when equipment is most likely to be handled by people who have never seen it before.
Audit the tags themselves periodically, not just the inventory. A sample walk that checks whether tags are present, legible, and scannable will find the adhesive failures and the faded print while there is still a record to correct them against.