If you create function-based index, oracle will act exactly the same as with any "normal" index - there will be no two sepparated indexes to support a single function based index, there is absolutely no need for oracle to do this internaly. And the internal structure of that function-based index will be the same as with any other B*Tree index.

And yes, of course function-based index will have an effect on inserts, updates or deletes, just like any other B*Tree index does (well, function-based index will have a slightly higher owerhead actually).