Why the scheme comes before the labels
Labels are printed once and read for years, by people who were not there when the cabling went in. If the scheme changes halfway through a build, the IDs on the first rows stop matching the drawings. Agree on the scheme, write it on one page, and give that page to every contractor before the first label is printed.
Build the ID from location parts
| Part | Example | Tip |
|---|---|---|
| Site | DC1 |
Only needed if you run more than one site. |
| Room or hall | H2 |
Use the name on the door. |
| Row | A |
Skip letters that look like digits, such as I and O. |
| Rack | 07 |
Two digits, with a leading zero. |
| Rack unit | U42 |
Decide once whether U1 is at the top or the bottom. |
| Port | P24 |
As numbered on the panel or device. |
Put together, a panel port reads H2-A07-U42-P24: hall 2, row A, rack 07, unit 42, port 24.
Name both ends
The simplest cable ID is the pair of locations it connects, for example A07-U42-P24 / B03-U40-P12. Anyone holding either end knows where the other one is. If the marker is too small for both, give the cable a short unique number and keep the map of both ends in your DCIM tool or cable schedule.
Rules that keep IDs readable
- Fixed length.
A07, notA7. IDs then sort in the right order and line up on the label. - One separator. Pick a hyphen or a dot and use it everywhere.
- No lookalikes. Avoid letters that read as digits.
- Short enough to print large. The shorter the ID, the bigger the text on a small marker.
- Location, not purpose.
WEB-SERVER-3is wrong the day the server is replaced.A07-U42is still true.
Where standards fit
Many operators follow the ANSI/TIA-606 administration standard. If your contract names it, use its identifier formats. The approach above uses the same principle, an ID built from where things are, not its exact format.
Mistakes that cost time later
- Running numbers only (cable 1847): useless without the spreadsheet.
- Each contractor with its own scheme.
- Labels at one end only.
- Handwritten IDs that fade or smudge.