A company wants to ensure that their critical database is highly available and can withstand the failure of an entire AWS Region. What should they implement?