Typically the model for ORM's that I've used goes something like this:
Program startup: Initialize ORM db connection While running: make/modify/delete domain objects, commit changes, rinse and repeat. Program shutdown: commit any uncommitted changes if necessary, disconnect ORM db connection
What if someone is sniffing network traffic though (and what if the application gets used over the internet)? Are the persistent objects in any way secured? I don't recall seeing any documentation for any ORMs talking about encryption.