Overview

Messaging systems depend on directories for a variety of needs. Fetching email addresses for users and knowing what server their mailboxes are located on is perhaps the most obvious, but directories also hold large amounts of configuration data, such as how servers connect together, permissions, and so on.

In the first generation of Exchange, the Directory Store (DS) held all of this data. The link between the DS and the other Exchange components, such as the Store and the MTA, that need to access directory data is simple, since all of the components reside on the same physical server. Thus, a network outage might prevent a first-generation Exchange server from contacting other servers to transfer email or stop users from connecting to mailboxes, but it will not stop the interaction between the different Exchange components. Because the DS is located on the same server, access to its data is also fast and reliable.

You cannot deny that having a local copy of the directory on every Exchange server has its advantages (such as always having a local directory available rather than having to depend on a network service), but it also implies a lot of replication traffic to keep all the copies of the directory synchronized. As the number of servers grows in an organization, the replication traffic expands dramatically. In addition, Microsoft designed the Exchange DS to be single purpose so other applications cannot use it. Exchange 2000 addressed these issues by supporting the AD and moving to...