Hi there,

I have a situation where several (potentially 200-300) accounts must be
sharing a single listener socket (happens to be a
well-known port address). We want to be able to control access, ie disable
and re-enable, by account. My intuition (because I
have pretty well zero knowledge in this area!) tells me that it can't be
done without either:

1)  Modifying the listener process (sometimes doable, but not always)
or
2)  Intercepting connection requests, emulating the protocol as far as
necessary to determine how the connection wants to use
the service, and then <<somehow>> handing off the connection to a copy of
the original server which is now listening on a
different port.

Neither of these are exactly making my day full of grins.

My questions --

a) Is my assessment accurate? and
b) Are there any guidelines in this area, or things that I have overlooked?


What would be nice would be a generalized solution, so we don't have to
hand-craft different solutions for each case. But at
the moment I don't know if this dream is going to work out....


Hoping for some enlightenment,
Tony Furnivall