Not sure about your environment, but now that I have end-users to deal with again, well, lets just say I have more grey hair now. (of what's left, that is) ;\)

One of the biggest issues I've encountered is some systems don't respond to ping, either because the netops team blocked ICMP at their remote router, the Windows firewall got turned on (yeah, we have GPOs to turn it off, but...), or the user installs some firewall like Black Ice or changes the McAfee settings and turns on the McAfee firewall. Our users are local admins - not my decision, but I have to live with it. Anyway, we've resorted to using XNET to check some service, like "UPS" - if it tells us anything other than "Remote machine not found", we're good to check the system and push updates, etc.

Bottom line, I'm very much in agreement with Richard about pull vs push. You could write a little routine to get the timestamp of \\server\service\message.ini. One quick burst on the net. Compare that with the timestamp with the last timestamp stored locally (.INI or registry to survive reboots) and display the message if needed. Your local routine to check and display could be a system service using SrvAny, and it could launch the local GUI to do the actual display once it detected the need. The service could determine and possibly retrieve the current message, minimizing the work done by the GUI.

With a service checking every 10-15 minutes, the network load would not be that great, nor concentrated. As a service, it determines the interval at startup, so the queries are spread out over a period of time. You could even wait a random amount of time between 10 and 15 mintes.

BTW - I've got several kix utilities running as services with SrvAny.

Glenn
_________________________
Actually I am a Rocket Scientist! \:D