A posting at Jeff Atwood’s Coding Horror discusses project names. There have been some famous project names throughout the years. Longhorn and Orcas should both be familiar to Windows developers.
Naming servers is a little different than naming projects. Once you name a server, you are probably stuck with that name for years and years. While you are also stuck with a project name, the project will (hopefully) be completed at some point in time.
People inside and possibly outside your department and organization may refer servers by name. So might software. DNS definitely alleviates some of this. However, changing the name of a server is never a great idea, so coming up with an adequate name from the beginning is always a good thing.
There are a few of schools of thought on naming computers that I have encountered in my career.
- Name a desktop computer based on the person using the computer.
This method usually ends up just being the person’s domain login. If your name is John Smith, your domain account is probably jsmith and your computer is probably named, you guessed it, jsmith. - Name a computer (server) based on its function.
Pre-Win2k domain days, it was not uncommon to have machines named something like PDB or BDC, for Primary Domain Controller and Backup Domain Controller respectively. Mail is a popular name for the mail server, etc. - Name a computer based on a random scheme.
One customer site I visited had machines like QB1Z2. This had no meaning other than the fact that no other machine on the network had the same name. - Name a computer based on location or department.
In very large organizations it can be handy to be able to *find* a computer, especially if you are the IT guy. A hospital that I visited (a customer) had machines like Triag001 and ER004. This allowed IT staff to quickly find machines. Where is Triag004? Well, it must be in triage! - Name a computer (or often a server) based on a predetermined naming list.
This method is usually used for servers. I have seen the Periodic Table of Elements used at two different jobs sites. Imagine a mail server named Helium, or a file server named Oxygen. At one site a server was named Upgradium. At one of those two sites just mentioned, all of the externally-facing machines were named after local lakes. I thought this was a pretty good idea for two reasons. First, it was easy to remember if a given server was in the DMZ or could be seen (attacked) by the outside world just by the name. Second, people outside of our organization (partner businesses, etc) probably did not have a knack for remembering elements, but they might remember the name of a local lake.
In the end, I think all of these except #3 are fine ideas. I personally like #5 for servers because it gives developers and IT staff something fun to do: come up with lists of names. Its always fun to refer to servers in this fashion as well.
