Data distribution – Apple Mac OS X Server (Administrator’s Guide) User Manual

Page 69

Advertising

Directory Services

Processes no longer need to know how and where administrative data is stored. Open
Directory gets the data for them. If a process needs the location of a user’s home directory,
the process simply has Open Directory retrieve the information. Open Directory finds the
requested information, and then returns it, insulating the process from the details of how the
information is stored. If you set up Open Directory to access administrative data in several
directory domains, Open Directory automatically consults them as needed.

Some of the data stored in a directory domain is identical to data stored in UNIX
configuration files. For example, the authentication attributes, home directory location, real
name, user ID, and group ID—all stored in the user records of a directory domain—have
corresponding entries in the standard /etc/passwd file. However, a directory domain stores
much additional data to support functions that are unique to Mac OS X, such as support for
managed clients and Apple Filing Protocol (AFP) directories.

Data Distribution

Another characteristic of UNIX configuration files is that the administrative data they contain
is available only to the computer on which they are stored. Each computer has its own UNIX
configuration files. With UNIX configuration files, each computer that someone wants to use
must have that person’s user account settings stored on it, and each computer must store
the account settings for every person who may want to use the computer. To set up a
computer’s network settings, the administrator needs to go to the computer and directly
enter the IP address and other information that identifies the computer on the network.

Similarly, when user or network information needs to be changed in UNIX configuration files,
the administrator must make the changes on the computer where the files reside. Some
changes, such as network settings, require the administrator to make the same changes on
multiple computers. This approach becomes unwieldy as networks grow in size and complexity.