Separation Model Questions...

I've been reading the posts on this topic and they are all very informative. I have a db that's getting large (at least large for us). We are planning on combining some nearby stores in the area into one "master" database. After having read some of the posts I'm starting to think the separation model may be a good option. There are a couple of things I don't quite understand though...

(1) If these files are separted, do the both reside on the server side? The GUI file isn't on the local "c" drive, is it? How would you do a mass / instant update if it is (unless you do it the way someone described and just e-mail a new GUI file to everyone)

(2) My business is 24 hours a day... with about 30 users, does the separation model have any noticable differences in speed if the GUI file has to constantly go retrieve data from the back end?

(3) Calculations and scripting... do they get done on the GUI side then?

I think the separtion model is good for one major reason... size. Our business collects pictures and other files and those are starting to add up and making the file HUGE. I really liked what one person said about that... if there is mass GUI change, they can just change that, and then when the dump the old and open the new, I guess it all just "re-connects" to the back end.

PS, isn't this just like what FMP6 was?

Thanks

V