Active Directory Sites
Active Directory uses the concept of sites to logically represent the physical network topology. Active Directory uses the site topology to route replication traffic efficiently. Site topology is also used to route queries and authentication requests.
Site topology is not related to the domain hierarchy! A single domain can appear in many sites, and a site can contain many domains as well (see Figure 3.8).
Figure 3.8 Sites versus domains in Active Directory.
Creating Sites
Microsoft considers a site a group of TCP/IP-based subnets connected by fast and reliable connections. So, in layman's terms, what does that mean? The best rule of thumb is to consider any networks with 10MBps or faster (that is, LAN speeds) a site. Sites that are separated by WAN links must be separate sites.
A site link is the connection between two or more sites (as its name implies). Site links are used by Active Directory to determine the amount of available bandwidth between two sites.
Site links consists of four factors:
Cost Active DirectoryUses the cost value of a site link to ascertain when a site link should be used for replication traffic. The cost values of various site links determine the route that replication traffic follows throughout the network. For example, sites that are connected via low-speed modems and dial-up lines have a higher cost than sites connected with high-speed network backbones. Costs are a good way of creating fault- tolerant site links. For example, if you have two site links between two sites and assign a higher cost to one of the links, replication will only occur across the link with the lower value. If that link fails, the link with the higher cost will be utilized.
Replication ScheduleThe replication schedule of a site link indicates when the site link is available to replication traffic. This can be used to force replication traffic to only occur during off-peak hours (such as late at night).
Replication IntervalThe replication interval value determines how often replication changes are requested from the other side of the site link. By default, the replication interval is three hours or 180 minutes. The minimum period is 15 minutes.
TransportThe transport determines the transport protocol used for replication traffic. Replication between sites can occur synchronously by RPC (Remote Procedure Calls) over the IP transport or asynchronously via SMTP (Simple Mail Transport Protocol) over IP transport.
TIP
RPC over IP should be used most of the time, unless, due to firewall or router configurations, RPC traffic is denied. In this case, SMTP over IP should be used.
It is important to note that site links using SMTP over IP will ignore any replication scheduling parameters. SMTP mail is transmitted according to the configuration of the e-mail infrastructure.
Clients and Servers Are Associated with Sites
When booting up, Windows 2000 clients query a domain controller for their respective domain. The domain controller analyzes the client's IP address and determines which site the client belongs to. The domain controller returns the site name to the client and the client caches the information. The client will use the site name to determine resources at the site.
KCC Uses the Site Topology to Create Replication Connections
The Knowledge Consistency Checker (KCC) uses the site topology to determine and create replication connections between domain controllers. Site link information and site boundaries are used to make these determinations.
Intrasite Versus Intersite Replication
There are two types of replication traffic in Active Directory, intrasite and intersite. Intrasite replication traffic is between domain controllers within the same site. Intersite replication traffic is between domain controllers in different sites. The KCC tunes intrasite replication to minimize replication latency, whereas it tunes intersite replication to minimize bandwidth usage.
Table 3.1 describes some of the differences in the two forms of replication traffic.
Table 3.1 Intrasite Versus Intersite Replication
Intrasite |
Intersite |
Traffic is uncompressed. |
Traffic is compressed (to save bandwidth). |
Replication partners notify each other when changes must be replicated (to reduce latency). |
Replication partners do not notify each other (to save bandwidth). |
Replication partners poll one another periodically. |
Replication partners poll one another during scheduled intervals only. |
RCP over IP transport only. |
RCP over IP or SMTP over IP transports. |
Replication connections can be created between any two domain controllers in the same site. |
Replication connections can only be created between bridgehead servers. A bridgehead server is designated by the KCC. A bridgehead server is a domain controller that has been designated to perform all intersite replication for a particular site. |
Site Topology Information Is Stored Within the Configuration Container
All information related to site topology (sites, site links, subnets, and so on) is stored in the configuration container. The configuration container is replicated to every domain controller in the entire forest. As a result, any change to the site topology is replicated to every domain controller in the forest, thus creating replication traffic.