Scott Nixon

Are you still using static IPs when you configure software or devices? Do you enter the hostname when configuring applications? Shame on you!

Systems Administrators can use DNS to design logical and simple naming conventions. Most organizations will name servers by location, function and include meta data about the system such as serial numbers and operating system. An example would be NY2K3WEB01. Looks great, logical description, easy to remember, and also provides information about the system.

Our fictional, NY2K3WEB01 is our web server running the help desk software.  We are implementing the new help desk software which is web based. Rather and sending users to http://ny2k3web01/ we create a CNAME DNS of helpdesk so they all users can easily remember and access the Help Desk web site.

helpdesk IN CNAME ny2k3web01.example.com

This is a fairly typical setup for web servers so why not do this for e-mail, LDAP, anti-virus, switches, mail relays, and network applications. When you configure that new network application create an Alias (CNAME) record of salescrm and then configured the application so all clients connect to the server with either the alias or the FQDN. This will make it so much easier for your IT staff to remember.

salescrm IN CNAME ny2k8app04.internal.example.com

So why would you do this? A brainy IT guy might say, “Once you configured it’s done so this is a waste of time.” If your server dies and it has to be moved to a different box then all the clients will need to be re-configured. Brainy guys says, “I would just use the same server name or IP address.” Sure you can but it’s bad form.

Rather than wasting precise time trying to clean the server out of Active Directory, DNS, Anti-Virus, and Patch Management, you could be restoring the server. Yes, I am being contrarian to the status quo in most IT shops because I’ve seen it work and save time.