OT: Thinking about the Playstation fiasco, not to mention Wikileaks and others; even though it is hard to prevent all hacking, even with stolen passwords, how hard could it be to have a system that will simply not allow any large data dumps to come from certain databases?

Even if they are accessed constantly internally, there have to be characteristic that would identify any non legitimate bulk data transfer, particularly if it comes from (or goes to) an outside address.