If the above makes sense to you, you are certifiably insane.

Yet, that is exactly what happens in a 'secure' computer network. A logged in individual who is active on the system gets periodically booted out. What is the rationale for this? Does it indicate that the user accounting within the system is flawed - that periodically the users get jumbled together so the network doesn't know who is who - that it is possible for someone to hijack the user's session unbeknown to the user or the network?

Going back to the scenario mentioned up top: there should be a rational reason for escorting an employee from the building when they were identified and authorized on entry. Likewise, there is no legitimate argument for timing out an active user on the system.

It may be that the reason there are timeouts for active users is that the infrastructure is not communicating the authentication information throughout the various components. If this is the case, the infrastructure should be reconfigured to have a single entry point where users are authenticated - a gateway. From that point, any resource that is touched in the infrastructure is passed the authentication information from the gateway. The resource then evaluates the access privileges based on the passed credentials without prompting the user.

It is also likely that activity by the user is not being reported back to the gateway. If that is the case, the user's session will appear inactive to the gateway and so the gateway times the user out. If this is the case, the infrastructure components should be required to post activity notices back to the gateway. That way an active user session will always be seen as active by the authentication controlling gateway. Assuming the gateway timeout is 2 hours, if a user works 24 hours a day, they would never be disconnected from the network. This is the way it should be.