At my current employer, we’ve done just that and here’s a breakdown of what we came up with for our Windows Server naming standard.

Our current standard name consists of seven sections or “Classifications” that together comprise the server name and it goes something like this:

CC | T | XXX | 123 | F | NN | O

• For the first classification, CC, we define the city or location based on a two letter code that has been standardized by my employer.
• The second classification, T defines the type of server whether it’s a physical piece of hardware or virtual machine. For example we use P for physical and E for an ESX virtual machine.
• The third classification, XXX defines a descriptor that could be the three letter initials for the software company who’s software is running on the server. Purely discretionary.
• The forth classification, 123 defines a numerical identifier and this section has been somewhat controversial for me and other co-workers. More on that later.
• The fifth classification, F defines the actual function of the server. So a domain controller would be designated with D, or a SQL server would use a Q for example.
• The sixth classification, NN is only used for cluster nodes and would indicate the node number as 01 or 02 for example.
• The seventh classification, O defines a remote administration card or heartbeat NIC for a cluster node. We use HP servers which include an integrated lights out card (or iLO) which we designate as R.

So based on the above information, a typical server name would be as follows:

BFPAAD001D which breaks down this way, Bum Fuck (BF) Acme Active Directory (AAD) server number 001 which is a domain controller (D). Since the server is an HP and has an iLO which has it’s own IP and DNS name, the iLO designation for this server would be BFAAD001DR with the R for remote card.

Now if you noticed, this post is also categorized as a “Rant” as well so here it is.

Overall the naming standard is pretty close to perfect, for our environment, but the one issue I have is that some want to name servers with a numeric designation starting with 000. For myself I don’t agree with this practice and a great example of this would be when you have 10 servers that are all identical in function, like a Citrix farm. The numeric designation for these servers should begin at 001 and end with 010. That makes sense. It’s like if someone wanted to borrow five dollars (in one dollar bills) from you and you started your count at zero (0). So you would count 0, 1, 2, 3, 4 and 5, there’s your five dollars. That does not make sense!

So my question. What standard do you use for your server names and would you use a 0 numeric identifier?