I'm wondering about how an entire planet can pretty much be one open scene and how a game engine can handle it. for example, world of warcraft. you can walk from one end of kalimdor to the other and go into every single city and building without a single loading screen. does that mean in their game engine they have everything laid out in one scene? if so, thats a lot of area to manage. is it possible to do that in a game engine like unity? is it hard for an engine to handle an entire map like that?
Yes, it maybe possible, however it probably is not practical. Even with Occlusion Culling, Mipmapping (Varies the resolution of textures based on distance), and LOD (level of detail, varies the complexity of objects based on distance) you are still limited about how much you can load into memory at once, and limited by reasonable estimates on what your customers have available in hardware. It gets even more complex with MMORPG's where the system needs to track the actions of hundreds of players, monsters, and treasure pools. You should check with the Unity forms and see if the engine can even support MMROPGs.
Skyrim, even if it is single player, gets close to that over world concept. They still load dungeons, and some special locations in their own environments, but the entirety of the surface world is free from loading screens. The in game map is really just a orbital camera that has it's Z value locked. What you see in the map is reflective to what actually happening in the game world.