I am having problems to create an Entity Framework Code-First mapping for the following sample database schema (in SQL Server):
Not a solution, but a workaround (*): a nice choice is to use single Id columns (as ), usually auto-incremented, and providing database integrity using foreign keys, unique indexes, etc. More complex data integrity could be achieved with triggers, so maybe you could be heading that way, but you might leave that to the application business logic level, unless the application is really data-focused. But since you're using Entity Framework, it's probably safe to assume that this is not your case...?
(*) as suggested by ivowiblo