Sorry to be repetitive, but we have quite a few internet socket connections through VPN, and I used to have the same issues with timeouts/broken connections until I reduced the UNIX (AIX) “keepalive” setting from the default 2 hours to 15 minutes. 30 minutes may work as well but I don’t think 15 mins is a problem re: resources. The engine’s keep alive pulse is really determined by the UNIX level keepalive setting. Now we never have timeout problems. Sometimes a remote system will still go down (ie. when they do backups or maintenance) where they need to cycle the interface server to bring the connection back online. But other than that (and other “system issues”) the connections stay up even with little/no activity.
