Sites and Replication
A site represents a region of uniformly good network access, which can be interpreted as being generally equivalent to local area network (LAN) connectivity. LAN connectivity assumes high, inexpensive bandwidth that allows similar and reliable network performance, regardless of which two computers in the site are communicating. This quality of connectivity does not indicate that all servers in the site must be on the same network segment nor that hop counts between all servers must be identical. Rather, it can be interpreted as the measure by which you know that if a large amount of data needed to be copied from one server to another, it would not matter to you which servers were involved. If you find that you are concerned about such situations, you might consider creating another site.
Replication within a site is driven by changes. The high speed and reliability of LAN connectivity lends itself to on-demand data transfer. When a change is applied to a specific replica, the replication engine is triggered. The replication engine waits for a configurable interval (by default, five minutes) and then notifies the first replication partner. Each additional partner is notified after a configurable delay (by default, 30 seconds).
Generation of the topology within a site is achieved by using the minimum number of connections possible. The default replication topology in a site is a bidirectional ring, with sufficient additional connections added to keep the number of hops between replication partners to three or less. To summarize, the default worst-case delay for propagation within a site is the maximum number of hops between a source and destination domain controller (3 hops) multiplied by the delay at each domain controller in the path (5 minutes), or 15 minutes for three hops.
Replication between sites is scheduled so that you can control replication costs and keep replication from overwhelming your communication links. Several factors contribute to the speed of replication between sites, including the nature of the physical connections and how expensive they are to use. Active Directory supports site connections in all speed ranges, from T3#160;links on the high end to dial-up lines on the low end. You use the speed of the connection between your sites to assign a cost to the communication link, and replication uses the cost to establish the least expensive route for replication traffic.
For information about planning networks, see "Determining Network Connectivity Strategies" in the Deployment Planning Guide .
Site Design with Replication in Mind
When you group a set of IP subnets into a site, you do so based on the fact that these subnets have high bandwidth, LAN connectivity, possibly involving hops through high-performance routers.
A single domain can span multiple sites, and a single site can contain multiple domains. Domain architecture should be constructed independently of site design — sites exist primarily, but not solely, to assist the domain controller Locator and the replication infrastructure. Although there are no absolute rules for deciding when to place two subnets in the same site, understanding how Active Directory uses site information can help you make an informed decision.
Active Directory uses site information in the following ways:
When a client requests a connection to a domain controller (for example, when logging on), sites are used to enable the client to connect to a domain controller with good connectivity whenever possible. Fast connections reduce network latency and conserve network bandwidth.
When the KCC configures replication connections between domain controllers, it creates more connections between domain controllers in the same site than between domain controllers in different sites. The results are lower replication latency within a site and less replication bandwidth between sites.
Replication messages between domain controllers within a site are uncompressed, which means that fewer CPU cycles are used on the domain controllers. Replication messages between domain controllers in different sites are compressed, which means that less network bandwidth is used.
Replication between domain controllers within a site is triggered by the arrival of updates, which thereby reduces replication latency within a site. Replication between domain controllers in different sites is performed on a schedule, which thereby conserves network bandwidth between sites.
Sites are not tied in any way to the Active Directory namespace that is used by the domain directory partitions. The name of a domain directory object does not reflect the site or sites in which the object is stored.
In the Exchange Server directory service, sites are tied to the namespace.
The hierarchy in the Sites container reflects object names within the configuration directory partition, where site names do appear in the distinguished names of objects in the Configuration container hierarchy, as seen in Active Directory Sites and Services.
For information about planning sites and site topology for Active Directory, see "Designing the Active Directory Structure" in the Deployment Planning Guide .
When you plan your sites, you decide what subnets provide the best connectivity for replication and add them to the same site. You can use Active Directory Sites and Services to create subnets, and then you can create a site and associate the subnets with the site. You create a site by creating a site object in Active Directory and then defining a set of one or more subnets that belong to the site.
The determination of whether a computer is in a site is that its IP address maps to a subnet object in Active Directory. In the default directory, there is no default subnet object, so potentially a computer can be in the forest but have an IP subnet that is not contained in any site. For private networks, you can specify the network addresses that are provided by the Internet Assigned Numbers Authority (IANA) By definition, that range covers all of the subnets for the organization. However, where several class B or class C addresses are assigned, there would necessarily be multiple subnet objects that all mapped to the same default site.
To accommodate this situation, use the following subnets:
For class B addresses, subnet 126.96.36.199/2 covers all class B addresses.
For class C addresses, subnet 192.0.0.0/3 covers all class C addresses.
For information about subnets and subnet masks, see "Introduction to TCP/IP" in the TCP/IP Core Networking Guide . For information about creating subnets and assigning subnet masks, see Windows 2000 Server Help. For information about designing your network, see "Determining Network Connectivity Strategies" in the Deployment Planning Guide . For information about designing sites for Active Directory, see "Designing the Active Directory Structure" in the Deployment Planning Guide .
When to Define a New Site
When you have slow links between network segments, it is recommended that you create two sites and place domain controllers into the sites according to the following general rules:
Deploy at least one Global Catalog per site.
Deploy DNS servers on a site level.
The first domain controller in the forest is designated automatically as a Global Catalog server. When you create additional sites, you can use Active Directory Sites and Services to select the Global Catalog option in the properties for the NTDS Settings object of the server that you want to be the Global Catalog. Having a Global Catalog server in each site improves search performance because searches do not have to cross site boundaries. In addition, a Global Catalog server is required for logging on to the domain; if a connection between sites is not available, logging on is not possible.
If a Global Catalog server is not available in one site but there is another Global Catalog server in a remote site, the server in the remote site can be used for the logon process. If no Global Catalog is available in any site, the logon process proceeds with cached logon information.
The availability of DNS directly affects the availability of Active Directory. Clients rely on DNS to be able to find a domain controller, and domain controllers rely on DNS to find other domain controllers. As a general rule, you configure at least one DNS server in every site.
When you create sites and put servers into the sites, you connect the sites with site links and configure the links to reflect the network characteristics in terms of how often you want replication to occur on the link.
For information about planning sites and about domain controller, DNS server, and Global Catalog server placement, see "Designing the Active Directory Structure" in the Deployment Planning Guide . For information about Global Catalog servers, see "Active Directory Logical Structure" in this book.
When you install Active Directory on the first domain controller in the site, an object named Default-First-Site-Name is created in the Sites container. The first domain controller is necessarily installed into this site. Subsequent domain controllers are either installed into the site of the source domain controller (assuming the IP address maps to the site) or installed directly into an existing site. When your first domain controller has been installed, you can rename Default-First-Site-Name to the name that you want to use for the site.
When you install Active Directory on subsequent servers, if alternative sites have been defined in Active Directory and the IP address of the installation computer matches an existing subnet in a defined site, the domain controller is added to that site. Otherwise, it is added to the site of the source domain controller.
For information about how sites are identified for new domain controllers, see "Active Directory Data Storage" in this book.