Re: Server naming conventions

[Steven Tardy <sjt5 () ITS MSSTATE EDU>]

a wise man long ago decided to decouple the server purpose from the server name.
physical servers get retasked confusing inventory and tracking and management.

physical servers use tree names...
  fig / elm / oak / catalpa / pine / cedar / ...
then DNS CNAMEs pointing the purpose to the host.
  smtps -=> catalpa

our sister group used to use rocks/mineral. oak is easier to spell than tsavorite. (:
other groups on campus use cartoon characters and movie characters.


today with virtual servers, virtual servers are named based on purpose.
... a VM wouldn't be retasked. you'd simply create a new VM.

distinct names makes things easier to remember:
  catalpa is part of smtps
as opposed to:
  srv## is part of smtps

can't remember what this server does?
we have a file on each server with it's purpose.
login and cat a file.


Daniel Woodruff wrote:
> What kinds of naming conventions do everyone follow when building new
> servers?
>
>
>
> Currently, our Windows hosts are named following the pattern 'its-w2ks#'
> or similar, where the # is the next in the sequence, and the names are
> published in DNS. What are the potential drawbacks or using a scheme
> like this? Do you think it is any better or worse from a security
> perspective than using something like 'its-oracle-1' which has the
> service right in the name? We're concerned about disclosing the purpose
> of the machine via its name, and are trying to get an idea of what other
> schools do for their machines. Thanks in advance.

--
Steven Tardy
Systems Programmer
Information Technology Infrastructure
Information Technology Services
Mississippi State University
sjt5 () its msstate edu