What are the differences between Windows NT 4 and Windows Server 2003?
In NT 4, there are primary domain controllers (PDCs) and backup
domain controllers (BDCs). Only one PDC can exist per domain. The
PDC hosts a read/write copy of the domain database, and the BDCs
each house a read-only version of the same information, thus
providing some measure of redundancy and infrastructure
scalability. But in the event of a loss of the PDC, you can't make
changes to the domain until you replace it. However, users can
continue to log in using a BDC. In Active Directory, with the
exception of specific roles (i.e., Flexible Single Master
Operations, or FSMO), all domain controllers function equally. The
system replicates changes made to read/write versions of the domain
database to all other domain controllers, resulting in a much more
robust and scalable infrastructure. A single domain controller does
serve the FSMO roles, but you can move these roles to any other
domain controller at will. Another key difference between NT and
Active Directory is the process by which a server becomes a domain
controller. In NT, once a server is a PDC or a BDC, only
reinstalling Windows can downgrade it to a member server (and vice
versa). But in Windows 2000 and Windows Server 2003, you use the
DCPromo utility to handle this process. This utility can upgrade
and downgrade servers to and from domain controllers and member
servers