Let me give you all a little background info. I started at my current job 11 months ago as a Systems Engineer for a bank. Not a huge bank, but decent sized -- $2.5 billion in assets.
This place was a MESS. ZERO server documentation, very little network documentation, carpeted server room with active sprinklers and no temperature or humidity control, VPN access granted with a keyfile only (no username or password required), using eDirectory for workstation authentication and THAT'S ALL meaning users had to remember about 5 other usernames and passwords.
The lack of network documentation is what this thread stems from. We're migrating to a virtual environment and whenever I p2v a server I keep the old one up with a new IP for a hot failover. Plus I've been setting up a central syslog server, Cacti for server and network monitoring, a new trouble ticket system, etc. which all run on Linux VMs. As you can see, knowing exactly what devices have what IP is ever changing and critical.
Our management subnet has devices all over the place with no rhyme or reason. They are literally strewn all over a class C.
This week I finally had enough of the lack of network documentation and decided to add that to my "hit list" (it's not my duty as a server guy but I want it done) and get something in place that is easy to use, pretty (to encourage the other IT staff to use it), and powerful enough to record not only our subnet but our satellite data center, our WAN subnets, MPLS subnets, etc. Again -- this shit was NOT being documented officially, someone would just add it to their own spreadsheet for personal reference if they remembered to do so.
So I set up TCPDB on a Debian VM and so far it's perfect. Took no time to set up and works great.