1.11 Sites

A site is a collection of IP subnets that share LAN-type connectivity. In fact, the best way to think of a site is to compare it to a LAN. Sites reflect locality all the systems that belong to the same site are close to each other and benefit from good bandwidth. Sites play a major role in replication of data.

When a Windows 2000 server creates a new domain, the AD creates the site Default-First-Site-Name and stores the DC there. All the DCs joining the domain are added to the default site. Systems continue to be added to the default site until a new site is explicitly created. A domain may span multiple sites (see Figure 1.7); however, multiple domains may belong to the same site. Sites are independent from the domains they belong to.

Figure 1.7: Steps for finding the closest site.

Sites are used for the following two roles:

1. During workstation logon to determine the closest DC. We will get into how workstations discover the closest DC in the next section.

2. During AD data replication to optimize the route and transport of replicated data depending on the DC sites.

In order to achieve these two roles, all DCs in the forest know about all the sites, which are stored in the configuration NC. Finding the closest DC at boot time or, more precisely, when the net logon service starts, DCs publish information about the domain and site they belong to in DNS using SRV records. As previously...