As I've posted before I'm not a networking expert but have ended up responsible for a small company network. The server runs Server 2003 Standard ( not R2) with SP2. This was an upgrade from server NT rather than a clean installation should that be significant. The workstations are all running XP Pro, most SP3 but a few old ones still SP2. However the issues happen with both systems running SP2 and SP3. We have two issues with workstations loosing connection to the server. Each workstation has various folders on the server as mapped drives. The first issue is that at exactly 18:00 GMT / 19:00 BST some, not all, workstations loose access. The issue got forgotten over the summer as nobody ever works until 19:00 but now an issue again as some are still working after 18:00. They end up saving their work locally then copying to the sever next day. If they rebooted they would get on the network again but of course just want to shut down and go home. Every user profile allows access 00:00 to 23:59:59 every day. The DHCP IP address lease duration is set to 16 hours so neither of these ought to be the reason. The second is some users loose access if they are away from their desk for a while. I can't recall the exact screens but have looked and as far as I can see the users should not automatically be disconnected after a time period of inactivity. After reading a website page I have changed the value of KeepConn to ffffffff I can not spot anything different in the user profiles on the server between the users affected and those who are not. Some of the affected users were profiles that existed on the NT server and the others created in Server 2003. Worst case while not solving the problem is there an application that could be loaded automatically on startup which would be invisible to the user and have negligible overhead that could be set to move/copy a file from the server then move/copy it back again every 5 - 10 minutes so the server sees the workstation as active? As far as I can see I can't create a batch file then run as a scheduled task more often than once a day. At least this would prove if either or both issues still occur when there is always data transfer every few minutes. Any ideas / items to check will be appreciated.