||This article may be too technical for most readers to understand. (November 2014)|
|This article needs additional citations for verification. (November 2014)|
An AD domain controller authenticates and authorizes all users and computers in a Windows domain type network—assigning and enforcing security policies for all computers and installing or updating software. For example, when a user logs into a computer that is part of a Windows domain, Active Directory checks the submitted password and determines whether the user is a system administrator or normal user.
- 1 History
- 2 Logical structure
- 3 Physical structure
- 4 Implementation
- 5 Database
- 6 Single server operations
- 7 Trusting
- 8 Unix integration
- 9 See also
- 10 References
- 11 External links
Active Directory, like many information-technology efforts, originated out of a democratization of design using Request for Comments or RFCs. The Internet Engineering Task Force (IETF), which oversees the RFC process, has accepted numerous RFCs initiated by widespread participants. Active Directory incorporates decades of communication technologies into the overarching Active Directory concept then makes improvements upon them.
For example, Lightweight Directory Access Protocol (LDAP), a long-standing directory technology, underpins Active Directory. Also X.500 directories and the Organizational Unit preceded the Active Directory concept that makes use of those methods. The LDAP concept began to emerge even before the founding of Microsoft in April 1975, with RFCs as early as 1971. RFCs contributing to LDAP include RFC 1823 (on the LDAP API, August 1995), RFC 2307, RFC 3062, and RFC 4533.
Microsoft previewed Active Directory in 1999, released it first with Windows 2000 Server edition, and revised it to extend functionality and improve administration in Windows Server 2003. Additional improvements came with Windows Server 2003 R2, Windows Server 2008, and Windows Server 2008 R2. With the release of the last, Microsoft renamed the domain controller role (see below) as Active Directory Domain Services (AD DS). It is also included in Windows Server 2012 and Windows Server 2012 R2.
As a directory service, an Active Directory instance consists of a database and corresponding executable code responsible for servicing requests and maintaining the database. The executable part, known as Directory System Agent, is a collection of Windows services and processes that run on Windows 2000 and later. Objects in Active Directory databases can be accessed via LDAP protocol, ADSI (a component object model interface), messaging API and Security Accounts Manager services.
An Active Directory structure is an arrangement of information about objects. The objects fall into two broad categories: resources (e.g., printers) and security principals (user or computer accounts and groups). Security principals are assigned unique security identifiers (SIDs).
Each object represents a single entity—whether a user, a computer, a printer, or a group—and its attributes. Certain objects can contain other objects. An object is uniquely identified by its name and has a set of attributes—the characteristics and information that the object represents— defined by a schema, which also determines the kinds of objects that can be stored in Active Directory.
The schema object lets administrators extend or modify the schema when necessary. However, because each schema object is integral to the definition of Active Directory objects, deactivating or changing these objects can fundamentally change or disrupt a deployment. Schema changes automatically propagate throughout the system. Once created, an object can only be deactivated—not deleted. Changing the schema usually requires planning.
Forests, trees, and domains
The Active Directory framework that holds the objects can be viewed at a number of levels. The forest, tree, and domain are the logical divisions in an Active Directory network.
Within a deployment, objects are grouped into domains. The objects for a single domain are stored in a single database (which can be replicated). Domains are identified by their DNS name structure, the namespace.
A domain is defined as a logical group of network objects (computers, users, devices) that share the same active directory database.
A tree is a collection of one or more domains and domain trees in a contiguous namespace, linked in a transitive trust hierarchy.
At the top of the structure is the forest. A forest is a collection of trees that share a common global catalog, directory schema, logical structure, and directory configuration. The forest represents the security boundary within which users, computers, groups, and other objects are accessible.
|Example of the geographical organizing of zones of interest within trees and domains.|
The objects held within a domain can be grouped into Organizational Units (OUs). OUs can provide hierarchy to a domain, ease its administration, and can resemble the organization's structure in managerial or geographical terms. OUs can contain other OUs—domains are containers in this sense. Microsoft recommends using OUs rather than domains for structure and to simplify the implementation of policies and administration. The OU is the recommended level at which to apply group policies, which are Active Directory objects formally named Group Policy Objects (GPOs), although policies can also be applied to domains or sites (see below). The OU is the level at which administrative powers are commonly delegated, but delegation can be performed on individual objects or attributes as well.
Organizational units are not mutually exclusive database; e.g. it is not possible to create user accounts with an identical username (sAMAccountName) in separate OUs, such as "fred.staff-ou.domain" and "fred.student-ou.domain", where "staff-ou" and "student-ou" are the OUs. This is so because sAMAccountName, a user object attribute, must be unique within the domain.  However, two users in different OUs can have the same Common Name (CN), the name under which they are stored in the directory itself.
In general the reason for this lack of allowance for duplicate names through hierarchical directory placement, is that Microsoft primarily relies on the principles of NetBIOS, which is a flat-file method of network object management that for Microsoft software, goes all the way back to Windows NT 3.1 and MS-DOS LAN Manager. Allowing for duplication of object names in the directory, or completely removing the use of NetBIOS names, would prevent backward compatibility with legacy software and equipment.
As the number of users in a domain increases, conventions such as "first initial, middle initial, last name" (Western order) or the reverse (Eastern order) fail for common family names like Li (李), Smith or Garcia. Workarounds include adding a digit to the end of the username. Alternatives include creating a separate ID system of unique employee/student id numbers to use as account names in place of actual user's names, and allowing users to nominate their preferred word sequence within an acceptable use policy.
Because duplicate usernames cannot exist within a domain, account name generation poses a significant challenge for large organizations that cannot be easily subdivided into separate domains, such as students in a public school system or university who must be able to use any computer across the network.
In Microsoft's Active Directory, OUs do not confer access permissions, and objects placed within OUs are not automatically assigned access privileges based on their containing OU. This is a design limitation specific to Active Directory. Other competing directories such as Novell NDS are able to assign access privileges through object placement within an OU.
Active Directory requires a separate step for an administrator to assign an object in an OU as a member of a group also within that OU. Relying on OU location alone to determine access permissions is unreliable, because the object may not have been assigned to the group object for that OU.
A common workaround for an Active Directory administrator is to write a custom PowerShell or Visual Basic script to automatically create and maintain a user group for each OU in their directory. The scripts are run periodically to update the group to match the OU's account membership, but are unable to instantly update the security groups anytime the directory changes, as occurs in competing directories where security is directly implemented into the directory itself. Such groups are known as Shadow Groups. Once created, these shadow groups are selectable in place of the OU in the administrative tools.
Microsoft refers to shadow groups in the Server 2008 Reference documentation, but does not explain how to create them. There are no built-in server methods or console snap-ins for managing shadow groups.
The division of an organization's information infrastructure into a hierarchy of one or more domains and top-level OUs is a key decision. Common models are by business unit, by geographical location, by IT Service, or by object type and hybrids of these. OUs should be structured primarily to facilitate administrative delegation, and secondarily, to facilitate group policy application. Although OUs form an administrative boundary, the only true security boundary is the forest itself and an administrator of any domain in the forest must be trusted across all domains in the forest.
The Active Directory database is organized in partitions, each holding specific object types and following a specific replication pattern. Microsoft often refers to these partitions as 'naming contexts'. The 'Schema' partition contains the definition of object classes and attributes within the Forest. The 'Configuration' partition contains information on the physical structure and configuration of the forest (such as the site topology). Both replicate to all domains in the Forest. The 'Domain' partition holds all objects created in that domain and replicates only within its domain.
Sites are physical (rather than logical) groupings defined by one or more IP subnets. AD also holds the definitions of connections, distinguishing low-speed (e.g., WAN, VPN) from high-speed (e.g., LAN) links. Site definitions are independent of the domain and OU structure and are common across the forest. Sites are used to control network traffic generated by replication and also to refer clients to the nearest domain controllers (DCs). Microsoft Exchange Server 2007 uses the site topology for mail routing. Policies can also be defined at the site level.
Physically, the Active Directory information is held on one or more peer domain controllers, replacing the NT PDC/BDC model. Each DC has a copy of the Active Directory. Servers joined to Active Directory that are not domain controllers are called Member Servers. A subset of objects in the domain partition replicate to domain controllers that are configured as global catalogs. Global catalog (GC) servers provide a global listing of all objects in the Forest. Global Catalog servers replicate to themselves all objects from all domains and hence, provide a global listing of objects in the forest. However, to minimize replication traffic and keep the GC's database small, only selected attributes of each object are replicated. This is called the partial attribute set (PAS). The PAS can be modified by modifying the schema and marking attributes for replication to the GC. Earlier versions of Windows used NetBIOS to communicate. Active Directory is fully integrated with DNS and requires TCP/IP—DNS. To be fully functional, the DNS server must support SRV resource records, also known as service records.
Active Directory synchronizes changes using multi-master replication. Replication by default is 'pull' rather than 'push', meaning that replicas pull changes from the server where the change was effected. The Knowledge Consistency Checker (KCC) creates a replication topology of site links using the defined sites to manage traffic. Intrasite replication is frequent and automatic as a result of change notification, which triggers peers to begin a pull replication cycle. Intersite replication intervals are typically less frequent and do not use change notification by default, although this is configurable and can be made identical to intrasite replication.
Each link can have a 'cost' (e.g., DS3, T1, ISDN etc.) and the KCC alters the site link topology accordingly. Replication may occur transitively through several site links on same-protocol site link bridges, if the cost is low, although KCC automatically costs a direct site-to-site link lower than transitive connections. Site-to-site replication can be configured to occur between a bridgehead server in each site, which then replicates the changes to other DCs within the site. Replication for Active Directory zones is automatically configured when DNS is activated in the domain based by site.
Replication of Active Directory uses Remote Procedure Calls (RPC) over IP (RPC/IP). Between Sites SMTP can be used for replication, but only for changes in the Schema, Configuration, or Partial Attribute Set (Global Catalog) GCs. SMTP cannot be used for replicating the default Domain partition.
In general, a network utilizing Active Directory has more than one licensed Windows server computer. Backup and restore of Active Directory is possible for a network with a single domain controller, but Microsoft recommends more than one domain controller to provide automatic failover protection of the directory. Domain controllers are also ideally single-purpose for directory operations only, and should not run any other software or role.
Certain Microsoft products such as SQL Server and Exchange can interfere with the operation of a domain controller, necessitating isolation of these products on additional Windows servers. Combining them can make configuration or troubleshooting of either the domain controller or the other installed software more difficult. A business intending to implement Active Directory is therefore recommended to purchase a number of Windows server licenses, to provide for at least two separate domain controllers, and optionally, additional domain controllers for performance or redundancy, a separate file server, a separate Exchange server, a separate SQL Server, and so forth to support the various server roles.
Physical hardware costs for the many separate servers can be reduced through the use of virtualization, although for proper failover protection, Microsoft recommends not running multiple virtualized domain controllers on the same physical hardware.
The Active-Directory database, the directory store, in Windows 2000 Server uses the JET Blue-based Extensible Storage Engine (ESE98) and is limited to 16 terabytes and 2 billion objects (but only 1 billion security principals) in each domain controller's database. Microsoft has created NTDS databases with more than 2 billion objects. (NT4's Security Account Manager could support no more than 40,000 objects). Called NTDS.DIT, it has two main tables: the data table and the link table. Windows Server 2003 added a third main table for security descriptor single instancing.
Single server operations
Flexible Single Master Operations Roles (FSMO, sometimes pronounced "fizz-mo") operations are also known as operations master roles. Although domain controllers allow simultaneous updates in multiple places, certain operations are supported only on a single server. These operations are performed using the roles listed below:
|Schema Master||1 per forest||Schema modifications|
|Domain Naming Master||1 per forest||Addition and removal of domains if present in root domain|
|PDC Emulator||1 per domain||Provides backwards compatibility for NT4 clients for PDC operations (like password changes). The PDC runs domain specific processes such as the Security Descriptor Propagator (SDP), and is the master time server within the domain. It also handles external trusts, the DFS consistency check, holds current passwords and manages all GPOs as default server.|
|RID Master||1 per domain||Allocates pools of unique identifiers to domain controllers for use when creating objects|
|Infrastructure Master||1 per domain/partition||Synchronizes cross-domain group membership changes. The infrastructure master should not be run on a global catalog server (GCS) unless all DCs are also GCs, or the environment consists of a single domain.|
To allow users in one domain to access resources in another, Active Directory uses trusts.
Trusts inside a forest are automatically created when domains are created. The forest sets the default boundaries of trust, and implicit, transitive trust is automatic for all domains within a forest.
- One-way trust
- One domain allows access to users on another domain, but the other domain does not allow access to users on the first domain.
- Two-way trust
- Two domains allow access to users on both domains.
- Trusted domain
- The domain that is trusted; whose users have access to the trusting domain.
- Transitive trust
- A trust that can extend beyond two domains to other trusted domains in the forest.
- Intransitive trust
- A one way trust that does not extend beyond two domains.
- Explicit trust
- A trust that an admin creates. It is not transitive and is one way only.
- Cross-link trust
- An explicit trust between domains in different trees or in the same tree when a descendant/ancestor (child/parent) relationship does not exist between the two domains.
- Joins two domains in different trees, transitive, one- or two-way.
- Forest trust
- Applies to the entire forest. Transitive, one- or two-way.
- Can be transitive or nontransitive (intransitive), one- or two-way.
- Connect to other forests or non-AD domains. Nontransitive, one- or two-way.
Windows Server 2003 introduced the forest root trust. This trust can be used to connect Windows Server 2003 forests if they are operating at the 2003 forest functional level. Authentication across this type of trust is Kerberos-based (as opposed to NTLM).
Forest trusts are transitive for all the domains within the trusted forests. However, forest trusts are not transitive between forests.
Example: Suppose that a two-way transitive forest trust exists between the forest root domains in Forest A and Forest B, and another two-way transitive forest trust exists between the forest root domains in Forest B and Forest C. Such a configuration lets users in Forest B access resources in any domain in either Forest A or Forest C, and users in Forest A or C can access resources in any domain in Forest B. However, it does not let users in Forest A access resources in Forest C, or vice versa. To let users in Forest A and Forest C share resources, a two-way transitive trust must exist between both forests.
Varying levels of interoperability with Active Directory can be achieved on most Unix-like operating systems (including Unix, Linux, Mac OS X or Java and Unix-based programs) through standards-compliant LDAP clients, but these systems usually do not interpret many attributes associated with Windows components, such as Group Policy and support for one-way trusts.
Third parties offer Active Directory integration for Unix-like platforms, including:
- Fox Technologies and the product FoxT ServerControl (software) implements AD Bridging capabilities that allows Unix-like systems to join Active Directory and enables the use of the Kerberos for authentication of users
- Centrify DirectControl (Centrify) – Active Directory-compatible centralized authentication and access control
- Centrify Express (Centrify) – A suite of free Active Directory-compliant services for centralized authentication, monitoring, file-sharing and remote access
- UNAB (Computer Associates)
- TrustBroker (CyberSafe Limited) – An implementation of Kerberos
- PowerBroker Identity Services, formerly Likewise (BeyondTrust, formerly Likewise Software) – Allows a non-Windows client to join Active Directory
- Quest Authentication Services (Now part of Dell) (Formerly, Quest, Vintela) - AD authentication, Group Policy management, User/Group Migration tools, Auditing and Reporting
- ADmitMac (Thursby Software Systems)
- Samba – Can act as a domain controller
The schema additions shipped with Windows Server 2003 R2 include attributes that map closely enough to RFC 2307 to be generally usable. The reference implementation of RFC 2307, nss_ldap and pam_ldap provided by PADL.com, support these attributes directly. The default schema for group membership complies with RFC 2307bis (proposed). Windows Server 2003 R2 includes a Microsoft Management Console snap-in that creates and edits the attributes.
An alternate option is to use another directory service such as 389 Directory Server (formerly Fedora Directory Server, FDS), ViewDS Identity Solutions - ViewDS v7.2 XML Enabled Directory or Sun Microsystems Sun Java System Directory Server, with the latter two both being able to perform two-way synchronization with AD and thus provide a "deflected" integration, as non-Windows clients authenticate to this while Windows Clients authenticate to AD. Another option is to use OpenLDAP with its translucent overlay, which can extend entries in any remote LDAP server with additional attributes stored in a local database. Clients pointed at the local database see entries containing both the remote and local attributes, while the remote database remains completely untouched.
- Active Directory Explorer
- AGDLP (implementing role based access controls using nested groups)
- Directory Services Restore Mode
- Flexible single master operation
- List of LDAP software
- Univention Corporate Server
- "Directory System Agent". MSDN Library. Microsoft. Retrieved 23 April 2014.
- Solomon, David A.; Russinovich, Mark (2005). "13". (4 ed.). Redmond, Washington: Microsoft Press. p. 840. ISBN 0-7356-1917-4. Missing or empty
- "Active Directory on a Windows Server 2003 Network". Active Directory Collection. Microsoft. 13 March 2003. Retrieved 25 December 2010.
- "The LDAP Application Program Interface". Retrieved 2013-11-26.
- Windows Server 2003: Active Directory Infrastructure. Microsoft Press. 2003. pp. 1–8–1–9.
- "Organizational Units". Distributed Systems Resource Kit (TechNet). Microsoft. 2011.
An organizational unit in Active Directory is analogous to a directory in the file system
- "sAMAccountName is always unique in a Windows domain… or is it?". Joeware. 4 January 2012. Retrieved 18 September 2013.
examples of how multiple AD objects can be created with the same sAMAccountName
- Microsoft Server 2008 Reference, discussing shadow groups used for fine-grained password policies: http://technet.microsoft.com/en-us/library/cc770394%28WS.10%29.aspx
- "Specifying Security and Administrative Boundaries". Microsoft Corporation. 23 January 2005.
However, service administrators have abilities that cross domain boundaries. For this reason, the forest is the ultimate security boundary, not the domain.
- Andreas Luther. "Active Directory Replication Traffic". Microsoft Corporation. Retrieved 26 May 2010.
The Active Directory is made up of one or more naming contexts or partitions.
- "Sites overview". Microsoft Corporation. 21 January 2005.
A site is a set of well-connected subnets.
- "Planning for domain controllers and member servers". Microsoft Corporation. 21 January 2005.
[...] member servers, [...] belong to a domain but do not contain a copy of the Active Directory data.
- "What Is the Global Catalog?". Microsoft Corporation. 10 December 2009.
[...] a domain controller can locate only the objects in its domain. [...] The global catalog provides the ability to locate objects from any domain [...]
- "Attributes Included in the Global Catalog". Microsoft Corporation. 26 August 2010.
The isMemberOfPartialAttributeSet attribute of an attributeSchema object is set to TRUE if the attribute is replicated to the global catalog. [...] When deciding whether or not to place an attribute in the global catalog remember that you are trading increased replication and increased disk storage on global catalog servers for, potentially, faster query performance.
- "Directory data store". Microsoft Corporation. 21 January 2005.
Active Directory uses four distinct directory partition types to store [...] data. Directory partitions contain domain, configuration, schema, and application data.
- "What Is the Active Directory Replication Model?". Microsoft Corporation. 28 March 2003.
Domain controllers request (pull) changes rather than send (push) changes that might not be needed.
- "What Is Active Directory Replication Topology?". Microsoft Corporation. 28 March 2003.
SMTP can be used to transport nondomain replication [...]
- "Active Directory Backup and Restore". TechNet. Microsoft. Retrieved 5 February 2014.
- "AD DS: All domains should have at least two functioning domain controllers for redundancy". TechNet. Microsoft. Retrieved 5 February 2014.
- Posey, Brien (23 August 2010). "10 tips for effective Active Directory design". TechRepublic. CBS Interactive. Retrieved 5 February 2014.
Whenever possible, your domain controllers should run on dedicated servers (physical or virtual).
- "You may encounter problems when installing SQL Server on a domain controller (Revision 3.0)". Support. Microsoft. 7 January 2013. Retrieved 5 February 2014.
- Degremont, Michel (30 Jun 2011). "Can I install SQL Server on a domain controller?". Microsoft SQL Server blog. Retrieved 5 February 2014.
For security and performance reasons, we recommend that you do not install a standalone SQL Server on a domain controller.
- "Installing Exchange on a domain controller is not recommended". TechNet. Microsoft. 22 March 2013. Retrieved 5 February 2014.
- "Security Considerations for a SQL Server Installation". TechNet. Microsoft. Retrieved 5 February 2014.
After SQL Server is installed on a computer, you cannot change the computer from a domain controller to a domain member. You must uninstall SQL Server before you change the host computer to a domain member.
- "Exchange Server Analyzer". TechNet. Microsoft. Retrieved 5 February 2014.
Running SQL Server on the same computer as a production Exchange mailbox server is not recommended.
- "Running Domain Controllers in Hyper-V". TechNet. Microsoft. Planning to Virtualize Domain Controllers. Retrieved 5 February 2014.
You should attempt to avoid creating potential single points of failure when you plan your virtual domain controller deployment.
- efleis (8 June 2006). "Large AD database? Probably not this large". Blogs.technet.com. Retrieved 20 November 2011.
- Active Directory Service Interfaces, Microsoft
- "Domain and Forest Trusts Technical Reference". Microsoft Corporation. 28 March 2003.
Trusts enable [...] authentication and [...] sharing resources across domains or forests
- "How Domain and Forest Trusts Work". Microsoft Corporation. 11 December 2012. Retrieved 29 January 2013.
Defines several kinds of trusts. (automatic, shortcut, forest, realm, external)
- server 2008 domain trust
- Edge, Charles S., Jr; Smith, Zack; Hunter, Beau (2009). "Chapter 3: Active Directory". Enterprise Mac Administrator's Guide. New York City: Apress. ISBN 978-1-4302-2443-3.
- "Samba4/Releases/4.0.0alpha13". SambaPeople. SAMBA Project. Archived from the original on 15 November 2010. Retrieved 29 November 2010.
- "The great DRS success!". SambaPeople. SAMBA Project. 5 October 2009. Archived from the original on 13 October 2009. Retrieved 2 November 2009.
- "RFC 2307bis". Retrieved 20 November 2011.
- "Active Directory Administration with Windows PowerShell". Microsoft. Retrieved 7 June 2011.
- "Using Scripts to Search Active Directory". Microsoft. Retrieved 22 May 2012.
- "ITAdminTools Perl Scripts Repository". ITAdminTools.com. Retrieved 22 May 2012.
- "Win32::OLE". Perl Open-Source Community. Retrieved 22 May 2012.
- "Free Active Directory Admin Tools from SolarWinds". MarketWire. January 2012.
|Wikiversity has learning materials about Active Directory|
- Microsoft Technet: White paper: Active Directory Architecture (Single technical document that gives an overview about Active Directory.)
- Microsoft Technet: Detailed description of Active Directory on Windows Server 2003
- Microsoft MSDN Library: [MS-ADTS]: Active Directory Technical Specification (part of the Microsoft Open Specification Promise)
- Active Directory Application Mode (ADAM)
- Microsoft MSDN: Active Directory Schema
- Microsoft TechNet: Understanding Schema
- Microsoft TechNet Magazine: Extending the Active Directory Schema