Forests, domains and organizational units are considered logical structure because they don’t follow any subnet or network boundary. The physical structure of the Active Directory, which uses subnet/network boundaries, consists of domain controllers and sites.
The computer that stores a replica (copy) of the account and security information of the domain and defines the domain is known as the domain controller. A Windows domain controller is a Windows server with an NTFS partition running Active Directory services. The directory data (account and security information) is stored in the NTDS.DIT file on an NTFS partition on the domain controller. Access to domain objects is controlled by access control lists (ACLs). ACLs contain the permissions associated with objects that control which users can gain access to an object and what type of access users can gain to the objects. Lastly, the domain controller manages user-domain interactions including user logon processes, authentication and directory searches.
Active Directory uses multi-master replication. This means that there is no master domain controller/primary domain controller as there was in Windows NT. Instead, all domain controllers store writable copies of the directory. When a change is made to the one of the domain controller, it is the job of the domain controller to replicate those changes to other domain controllers within the same domain within a short period of time. By adding a domain controller to a domain, the server is automatically configured for replication.
A site is one or more IP subnets connected by a high-speed link (128 Kbps or higher), typically defined by geographical locations. Sites are based on IP subnets of which any subnet can only belong to one site. Multiple subnets can be assigned to a single site. When a user logs on, Active Directory clients locate an Active Directory server in the same site as the user.
When an update occurs on a domain controller, the replication engine waits for a configurable interval, which is five minutes by default. It then sends a notification message to the first replication partner, informing it of the change. Each additional direct partner is notified after a configurable delay, which is 30 seconds by default. Items that are security-sensitive are immediately replicated and partners are notified immediately. If no changes occur during a configuration period, which is one hour by default, a domain controller initiates replication with its replication partners to ensure that no changes from the originating domain controller were missed.
One reason to use sites is to control replication traffic. When a domain controller replicates to another domain controller in a site, replication information is done without being scheduled. Urgent changes such as password changes, account lockout policy changes, freshly locked accounts and domain password policies are replicated immediately. In addition, the replication data sent is sent uncompressed, which keeps the domain controller processing down to a minimum.
The default replication pattern used by the Active Directory is optimized for a single location or site with high-speed connectivity. However if your network includes multiple locations or sites, the replication pattern would have to be configured so that a domain controller will replicate all changes to the domain controllers within a site and have a single domain controller to replicate across a slower WAN link to the other sites.
A bridgehead server is a single server located in each site that is designated to perform site-to-site replication. Bridgehead servers are designated automatically or they can be assigned manually by an administrator. The link between bridgehead servers are assigned schedules of what times of day the link is available to carry replication traffic. The replication interval indicates how often the bridgehead servers poll the other side of the link for replication changes.
The first site in Active Directory is called “Default-First-Site-Name", which is created automatically for the administrator. This site is a member of the default site link called "DEFAULTIPSITELINK", which is also created automatically for the administrator. If the administrator creates two additional sites ("site1" and "site2" for example), the administrator must define a Site Link that each site is going to be a member of before they can be written to Active Directory.
By default, all site links are bridged or transitive; that is, all site links for a specific transport implicitly belong to a single site link bridge for that transport. If your IP network is not fully routed IP network, you do not need to configure any site link bridges. Replication within sites requires little or no planning because it is fully automatic. However, when you have multiple sites, you should use the following steps to optimize Active Directory synchronization traffic:
Site Replication
Inter-site Replication
As mentioned before, you should have more than one controller for fault tolerance and for faster authentication. Placing domain controllers and DNS Servers at the different physical sites greatly improves the performance of the directory in a wide area network (WAN) since authentication will be done locally and not through a slow WAN link. Of course, having multiple servers at a site will offer fault tolerance and load balancing for requests. In addition, don’t forget to have multiple global catalog servers.
To create a site
To create a site link
Note: If you create a site link that uses SMTP, you must have an enterprise certification authority (Enterprise CA) available and SMTP must be installed on all domain controllers that will use the site link.
To create a site link bridge
Note: If you have enabled Bridge all site links, this procedure is redundant and will have no effect.
Patrick Regan has been a PC technician, network administrator/engineer, design architect and security analyst for the past 16 years after graduating with a bachelor's degree in physics from the University of Akron. He is currently a senior network engineer at Pacific Coast Companies, supporting a large enterprise network. He holds many certifications including the Microsoft MCSE, MCSA, MCT; CompTIA's A+, Network+, Server+, Linux+, Security+ and CTT+; Cisco CCNA; and Novell's CNE and CWNP Certified Wireless Network Administrator (CWNA).
Over the last several years, he has written several textbooks, including Troubleshooting the PC, Networking with Windows 2000 and 2003, Linux, Local Area Networks, Wide Area Networks and the Acing Series (Acing the A+, Acing the Network+, Acing the Security+ and Acing theLinux+). He has also co-authored the ExamCram 70-290 MCSA/MCSE Managing Linux+ and Maintaining a Microsoft Windows Server 2003 Environment, 2nd Edition and authored the ExamCram 70-620 Microsoft Windows Vista, Configuration. He is currently writing the study guides for the A+ certification exams for Cisco Press and the ExamCram 70-643 Windows Server 2008 Applications Platform Configuration.
|
|