Just the perl stuff, right? I wrote and I'm maintaining a suite of PERL s‎crip‎ts to properly manage a client/server environment for CAD/PLM infrastructure. Server side: I have a PERL s‎crip‎t called at (AIX) server boot that load a (PERL storable) config file from the network to find out what this server is supposed to run (about 10 different services possible: Apache, orbix, other commercial apps). A bunch of s‎crip‎ts allow me to monitor services, stop them, start them, tail log files… Client side: I have a PERL/TK starter bar, used by Win and AIX users where they pick which instance they want to point to (QA, training, production) and which app they want to run. Based on user choice and configuration file (same file as for server) the starter is starting apps on clients pointing them to the right server (includes load balancing for some apps). In the middle: I have developed an administration console in PERL/TK that generates the configuration file, by modifying this central config file, both client and server side know how they should connect together, it’s a big improvement compare to standard setup where you have to go modify multiple files in multiple places to configure your client server environment. This stuff is used at 6 different sites worldwide for 25 servers and 500 users.