Label Scheme for Rack Mounted Equipment

Question

Given a lot of the users on here have vast experience, what is the best way to label individual cables for equipment mounted in a rack?

I was thinking something like:

Port Number of Switch | U-Number | Server Interface Number (such as Rc1-InterfaceID) on btoh ends of the cable but wanted to see what more experienced folks were using.

There is no single *best way*, it just depends on too many factors. Think about what you want documented and what makes sense in your environment and then establish this scheme. — Sven, Sep 19 '12 at 14:28

score 3 · Accepted Answer · answered Sep 19 '12 at 14:33

It ultimately depends on why you want the labels.

For me the most important thing is so that when something needs plugging back in, you don't have to trace the cables with your fingers - impossible in some circumstances. Your scheme would facilitate that, although you probably want to include:

Device/Patch ID (could be a server, switch, other network device or just a patch panel)
Interface ID (specifically which port the cable should be plugged into)

Include details of both ends of the cable at both ends, since sometimes you want to know where the far end of the cable goes!

I would say it is useful to use the physical labelling on the devices themselves, and then keep documentation of how that maps through in the specific system (e.g. NIC on a server may be labelled A, but in the OS it is eth0). Where you have multiple interface cards with multiple ports, you should agree and document a naming convention.

If you use VLANs you might want to include VLAN ID, or you might want to use different colour cables for different things.

I would avoid any tags that snag, and definitely anything that falls off easily.

So what kind of documentation do you use to match your physical labeling? Pictures of the back of the servers/switches/routers with some sort of color coding for each port to say which cables are going where? Not sure the best way to create a matching document — What'sTheStoryWishBone, Sep 20 '12 at 12:14

score 1 · Answer 2 · answered Sep 19 '12 at 15:21

As SvenW points out there is no single best way.

The big thing I've found to be advantageous is to not label cables using server name, port number, etc. These items become inherently inaccurate over time. It takes a long tome to physically relabel cables. It is much faster to update a document. As a result my preference is to label the cable on both ends with a unique identifier (two letters and two digits is my favorite, but anything works really) and call it good. The other information (port assignments, servers, etc) are maintained in separate document, spreadsheet, etc.

I'm also a fan of color coding cables if / where applicable. If for example you have ESXi servers which are cabled into two network switches for redundancy as well as into an iSCSI SAN I'd color code Network 1 green, 2 blue SAN 1 white, 2 grey.

Label Scheme for Rack Mounted Equipment

2 Answers2