The concept of a site is to represent physical locations in a wide-area network. Site definitions support the control of replication. Site knowledge is also used for logon location, printer location, and other uses. Sites group domain controllers that are physically located in well-connected areas.

• Intrasite replication is automatically configured to replicate between replication partners every 5 minutes. Because there is a maximum of 3 hops between domain controllers, replication is completed within 15 minutes. A service called KCC Knowledge Consistency Checker automatically determines the replication topology. The replication topology specifies which DCs exchange information (replication partners) such that there are no more that 3 hops between any two servers.
• Intersite replication is configurable and schedulable.

New York site		Los Angeles site LA
çintrasite replicationè	çintersite replicationè	çintrasite replicationè

Site information is used for

• Controlling the replication topology
• Scheduling replication
• Locating services within a site

Sites and Services Manager

Sites are created with the Sites and Services Manager. Rather than designing some new way of defining a site, the Active Directory uses the existing TCP/IP subnet structure that is already defined for routers to find various sites. The Active Directory sites do not need to follow the subnet structure exactly. Site subnet masks can be adjusted to combine or divide the TCP/IP subnets. For example, a site defined with the address 209.47.184.0 and mask 255.255.255.0 defines the range of addresses 209.47.184.1 to 209.47.184.254 as being in the same site.

Each site is associated with a subnet object which is defined by a IP network address and a subnet mask.  All computers within the subnet address range are considered to be at the same site. Site definitions normally match the physical subnets, but they can be specified to combine or divide physical subnets.

Each site connects to other sites with a site link which has configuration information that controls replication. 

The site link configuration shown in the following dialogs includes

• Link cost
• Replication frequency
• Replication Schedule

The replication schedule is easily controlled by modifying the matrix shown in the following dialog.

How does a client determine its site?

Administrators are not required to configure the site for each client. Portable computers often change their site locations and they need to automatically determine their new location without manual administrative effort. This problem also exists and already has solutions in the area of TCP/IP configuration. When portable computers change locations, they need to have a new IP address assigned. This can be done manually, but is most often done automatically by DHCP servers.

When a client computer starts, it will connect to any domain controller in its domain. The domain controller looks at the client IP address and compares it to the site definitions stored in its AD configuration name context. If the domain controller is in the same site, it will continue to provide services to the client. If the domain controller is in a different site, it will redirect the client to a domain controller in that site.

		è  I need to connect to a DC, here is my IP address  è ç  You are in my site so I will look after you  ç
		èI need to connect to a DC, here is my IP addressè çYou are in a different site, connect to a server in your siteç

The site location is remembered in the registry in the key
    HKLM\System\CurrentControlSet\Services\Netlogon\Parameters

• Value DynamicSiteName records the last known site location
• Value SiteName specifies a site location that will override the dynamic discovery