Having Fun at Layer 1

So, we’re planning a new space for eWeek Labs, and despite being the new kid in school (or perhaps because of it) I was tapped to draft a proposed specification for the network cabling and related infrastructure. I’ve spent the best part of the last three or four days working on drafts and have finally come up with a document that ought to satisfy the most literal of contractors.

This isn’t my first rodeo by a long shot; depending on how one does the math, this is either the fourth, fifth or sixth time I’ve been somewhat or totally responsible for planning a lab space or server room. I’ve racked my brains trying to remember all the things that went wrong with earlier installations, and I’m very lucky that on this proposal, I only have to worry about one type of cabling.

Since I don’t have to consider power or HVAC requirements – at least, not so far – this project has been an easy job. Things certainly have changed for the better in the last twenty years; back then, I was wrestling with serial terminals, IBM Type 1 STP, and oddball hardware that didn’t use standard cable, connectors or pinouts.

My proposal covers everything from cable trays and conduit, to grounding and bonding, to how cable and patch panels should be labeled. The good news is that our runs are fairly short throughout the space, so I can get away with specifying Category 6 UTP instead of the much more expensive Category 6a.

If you have suggestions for how to make the physical layer work for you, instead of against you, please feel free to share them. Likewise, we’d love to read your horror stories of cabling jobs that went horribly wrong.