1 Introduction

Active Directory® is a directory service (DS). Directory services can be used to provide a central store for identity and account information as well as storage of information for other systems and applications. Use of the Active Directory system is appropriate when there is a requirement for a DS. It is also appropriate when building another system that has a dependency on the Active Directory protocols. An example of such a system is the Group Policy system, which is described in the Group Policy Protocols Overview document [MS-GPOD].

This document describes the member protocols that comprise the Active Directory system. It also describes the abstract state that is shared between the system's protocols. This document is intended for anyone who plans to implement the Active Directory system because it provides a high-level introduction to the functionality of the system and also describes the protocols that an implementation of the Active Directory system supports.

This document does not duplicate or replace the content of the protocol specifications that describe the individual protocols in the Active Directory system. An implementer has to refer to those Technical Documents (TDs) for information about each protocol. Additionally, the Active Directory Technical Specification [MS-ADTS] contains vital information about the behavior of the DS, such as the state model and processing rules, that is essential to the correct functioning of the system.

A DS is a service that stores and organizes directory objects in a centralized, hierarchical data store. This hierarchical organization of objects is called the directory. A directory object is an object that contains one or more attributes. Each attribute can have one or more values. Directory objects are identified by a name that is unique among all directory objects in the DS. The directory objects are organized in a hierarchical manner with regards to other directory objects. For example, a DS might have a container directory object named Users, the contents of which (referred to as child directory objects) are containers named for each logical division of users; for example, Accounting Department, Human Resources Department, Engineering Department, and so on. The contents of each of these containers, in turn, could be user objects, each of which represents one individual user and contains attributes that store information about that user, such as the user name, password, or telephone number. The following diagram shows this example.

Example of directory organization

Figure 1: Example of directory organization

The Active Directory system can operate in two distinct modes: as Active Directory Lightweight Directory Services (AD LDS) and as Active Directory Domain Services (AD DS).

AD LDS consists of a directory service that is accessible via the Lightweight Directory Access Protocol (LDAP) versions 2 and 3. AD LDS is primarily intended for use by application software as a storage mechanism. Note that information that is applicable to AD LDS on applicable Windows Server releases is also generally applicable to AD LDS on Windows clients. For more information, see [MS-ADTS] section 1.

AD DS is also accessible via LDAP versions 2 and 3, but it extends the basic DS to include additional capabilities, such as the ability to host domain naming contexts (domain NCs), and additional protocols. This permits AD DS to store the account information for the users of a computer network. The collection of accounts that is stored in AD DS is referred to as a domain. Such account storage is a vital function of the Active Directory system, and in particular of AD DS. However, the Active Directory system is not limited to storing such information. Any information that can be represented as a collection of attribute/value pairs, including the possibility of multivalued attributes, can be modeled as a directory object and can be stored in the Active Directory system.

Except where noted otherwise, information in this document applies to both AD DS and AD LDS. The Active Directory system encompasses both AD DS and AD LDS.

Physically, the Active Directory system consists of one or more computer servers that run a directory service. In the case of both AD DS and AD LDS, these computers are referred to as domain controllers (DCs). Even though the directory service can be running on multiple computers, these computers replicate the contents of the directory so that a client sees a consistent view of the directory no matter which directory server or DC it communicates with. The network protocols that perform this replication are described in [MS-DRSR], [MS-NRPC] section 3.6, and [MS-SRPL].<1>

Note This document, like [MS-ADTS] and [MS-DRSR], uses the term "domain controller" to refer to a DS that runs as either AD DS or AD LDS. Both AD DS and AD LDS are considered to be directory services.

Directory objects can be created and deleted in the directory. Subsequent to its creation, the contents of a directory object can be modified by adding or removing attributes and their values, or by changing the values of existing attributes. Clients can retrieve the contents of directory objects either by reading the attributes of a specific object or by querying for any objects that match client-specified criteria.

The core protocol that clients use to perform operations is LDAP version 3 as described in [MS-ADTS] which is preferred over version 2. Clients can also communicate with the Active Directory system by using the network protocols described in [MS-DRSR], [MS-SRPL], [MS-SAMR], [MS-LSAD], [MS-LSAT], and [MS-DSSP], as well as the Web Service protocols described in [MS-ADDM], [MS-WSTIM], [MS-ADCAP], [MS-WSDS], and [MS-WSPELD]. The set of protocols that the Active Directory system supports depends on whether the system is running as AD DS or AD LDS (see section 2.8).

This document provides a description of the client-server functionality of the Active Directory system and additional specific benefits that a client realizes from being associated with an AD DS directory service (that is, "joined to a domain") that are described later in this document. This document does not describe identity and security concepts that are defined as part of the Windows Protocols Overview Document [MS-WPO].

A directory service is a service that stores and organizes directory objects in a centralized, hierarchical data store. A central concept in the directory service is the directory tree. A directory tree is an arrangement of directory objects into a tree structure. Each directory object has exactly one parent directory object, except for the object that serves as the root of the tree and has no parent. Each directory object can have zero or more child objects. Each directory object is assigned a name (the relative distinguished name (RDN)) that is unique among its sibling objects. Each directory object can be uniquely identified from among all the other objects in the directory service by its distinguished name (DN), which is formed by concatenating the RDNs of the directory objects along the path from the root of the tree to the specific object. For more information, see [MSDN-DomainTrees].

Domain Interaction

AD DS implements one or more domains within a forest. A domain provides a number of services to its clients, primarily related to security and management. The security principals of the domain are all available from the AD DS domain controller; therefore, the domain serves as the primary source of identity for the clients of the domain. The domain, through the relevant security protocols, provides the basis for authentication within the domain, allowing principals within the domain to establish authenticated connections with each other. During the authentication process, the domain provides authorization information in the form of additional identities that represent groups, enabling authorization decisions to be made.

AD DS stores directory data and manages communication between users and domains. This includes user logon processes, authentication, and directory searches. AD DS provides a distributed database that stores and manages information about network resources and application-specific data from directory-enabled applications. Administrators can use AD DS to organize elements of a network, such as users, computers, and other devices, into a hierarchical containment structure. The hierarchical containment structure includes the Active Directory forest, domains in the forest, and organizational units (OUs) in each domain.

The Active Directory system that runs as the AD DS service on an AD DS domain controller provides certain services and benefits to domain-joined clients. These benefits include support for the Kerberos authentication protocol, which domain-joined clients can use to authenticate to the AD DS domain controller and to each other. The benefits also include certificate autoenrollment, which automatically deploys certificates to domain-joined computers and to users whose accounts are stored in the directory service, and automatic deployment of administrator-configured policy settings.

Many network-related operations depend on domains in order to complete various tasks. This document describes some of these tasks, including:

  • Locating a domain controller using DNS and NetBIOS.

  • Joining a domain by creating an account via the Security Account Manager remote procedure call (RPC) protocol [MS-SAMR].

  • Joining a domain by creating an account via LDAP.

  • Removing a domain member.

This document includes protocols that are used to communicate with a domain controller and maintain state. It also includes protocols that are used to augment authentication and authorization actions, and protocols that are used to interact with domain controllers.

The domain controller serves a central role in an enterprise network by functioning as the root of authority for sets of users and computers. A domain controller aggregates functionality that relates to identity management, authentication, authorization, and other management policy. Clients of the domain functionality in turn rely on the domain controller to establish secure communication, authorize requests, and apply policy. A client of the domain can itself be a server of some other role, for example, a file server that is handling the file storage requirements of other client workstations.

Directory Replication

Active Directory is a distributed directory service that stores objects that represent real-world entities, such as users, computers, services, and network resources. Objects in the directory are distributed among all domain controllers in a forest, and all domain controllers can be updated directly. Active Directory replication is the process by which the changes that originate on one domain controller are automatically transferred to other domain controllers that store the same data.

The Active Directory replication model ensures that Active Directory data on one domain controller will eventually converge with the replica of the same data on other domain controllers in the same domain. The Active Directory replication model determines how changes to Active Directory data are propagated and tracked automatically between domain controllers. The replication model allows directory data on each domain controller to be updated directly. Each domain controller maintains replication metadata that indicates the update status both for itself and relative to other domain controllers. In addition, the replication model allows each domain controller to request (or pull) only the changes that need to be replicated and to forward changes to other domain controllers that require them.

When a change is made to an object in a directory partition, the value of the changed attribute or attributes is updated on all domain controllers that store a replica of the same directory partition. Domain controllers communicate data updates automatically through Active Directory replication. Their communication about updates is always specific to a single directory partition at a time.

A domain that is run by AD DS can consist of many partitions or naming contexts (NCs). The DN of an object includes enough information to locate a replica of the partition that holds the object. The global catalog (GC) contains a partial replica of every NC in the directory. It also contains the schema and configuration naming contexts (config NCs). This means that the GC holds a replica of every object in the directory but with only a small number of their attributes. The attributes in the GC are those that are most frequently used in search operations, such as a user's first and last names or logon names, and those required to locate the full replica of the object.

Active Directory objects are instances of schema-defined classes, which consist of named sets of attributes. Schema definitions determine whether an attribute can be administratively changed. Attributes that cannot be changed are never updated and therefore never replicated. However, most Active Directory objects have attribute values that can be updated.

Replication within a site occurs as a response to changes. On its NTDS Settings object, the source domain controller stores a repsTo attribute that lists all servers in the same site that pull replication from it. The Knowledge Consistency Checker (KCC) updates these attributes, as described later in this section.

When a change occurs on a source domain controller, it notifies its destination replication partner, prompting the destination domain controller to request the changes from the source domain controller. The source domain controller either responds to the change request with a replication operation or places the request in a queue if requests are already pending. Each domain controller has a queue of pending replication operations that are processed one at a time.

Directory Replication Service (DRS) Remote Protocol [MS-DRSR] is an RPC protocol for replication and management of data in Active Directory. Domain controllers use the Security Account Manager (SAM) Remote Protocol (Server-to-Server) [MS-SAMS] to forward time-critical database changes to the primary domain controller (PDC), and to forward time-critical database changes from a read-only domain controller (RODC) to a writable NC replica within the same domain outside the normal replication protocol. This protocol is used only between Active Directory servers in the same domain. Beginning with Windows Server 2008 operating system, this protocol was extended to forward certain write operations that are not time critical from an RODC to a writable NC replica. The SAMS protocol addresses the requirement to propagate a subset of database changes to the PDC more quickly than the Directory Replication Service (DRS) Remote Protocol. This rapid propagation is used for sensitive information when the delay imposed by standard Active Directory replication creates an unwelcome burden on the user or creates a risk to the enterprise. An example of the former is a password change operation; if the password is not made available rapidly, a user can experience unpredictable authentication failures when the new password is tried against domain controllers that have not yet replicated it. An example of the latter is when an account is locked out due to multiple password failures; the lockout condition, and, equally important, the lockout-cleared condition, has to be propagated rapidly throughout the domain.

Replication can be event-driven or schedule-driven as explained in [MS-ADTS] section

Knowledge Consistency Checker (KCC)

A domain controller that runs Active Directory is part of a distributed system that performs replication. The KCC is a component that reduces the administrative burden of maintaining a functioning replication topology. The KCC ensures that a replication path exists between the same NCs that are present in different DCs. The KCC is explained in [MS-ADTS] sections and 6.2. The KCC helps administrators build a replication topology that incurs minimal cost. This cost is defined by the administrator as explained in [MS-ADTS] section

FSMO Roles

Each DC accepts originating updates for most attributes of most objects within its writable NC replicas. However, certain updates are accepted only if the DC is the single designated "master" DC for the update. This mechanism is called flexible single master operation (FSMO).

If some or all of the updates to an object are single-mastered, that object belongs to a defined set of objects. [MS-DRSR] section (GetReplScope) specifies these sets, which are called FSMO roles. Each FSMO role is applicable to a certain scope: either domain-wide, or forest-wide. The domain-wide FSMO roles include the Infrastructure Master FSMO, the Rid Master FSMO, and the PDC Emulator FSMO. The forest-wide FSMO roles include the Domain Naming FSMO and the Schema Master FSMO. There are no FSMO roles that apply strictly to application NCs.

Because a server that is operating as AD LDS does not host domain NCs, it cannot own any of the three domain-specific FSMO roles. It can own the Schema Master FSMO and Domain Naming FSMO roles.

In a given NC, each FSMO role is represented by an object. [MS-DRSR] section (GetReplScope) specifies these objects, which are called FSMO role objects.

The fSMORoleOwner attribute of each FSMO role object is an object reference to the nTDSDSA object of the DC that owns the role; that is, the DC that performs updates to objects in the role. Information about nTDSDSA objects and how they represent DCs are specified in [MS-ADTS] section 6.1.

An originating update to an object within a FSMO role generates an LDAP referral if the DC that receives the request cannot perform the update; the referral is to the DC represented by the nTDSDSA object referenced by the FSMO role object's fSMORoleOwner attribute on the DC that received the request.

The processing of updates affected by FSMO roles is fully specified in [MS-ADTS] section

The IDL_DRSGetNCChanges method ([MS-DRSR] section 4.1.10) makes an originating update to the fSMORoleOwner attribute of a FSMO role object while preserving single-mastering of updates to the FSMO role. The ability to update the fSMORoleOwner attribute in this way is exposed through LDAP as the root DSE updates becomeDomainMaster, becomeInfrastructureMaster, becomePdc, becomePdcWithCheckPoint, becomeRidMaster, and becomeSchemaMaster, as specified in [MS-ADTS] section

Reading the rootDSE attribute valid FSMOs on a DC returns the set of all FSMO roles (represented as FSMO role objects) that the DC will update, as specified in [MS-ADTS] section

Active Directory Trust Management

Active Directory domains rarely exist in isolation. Many Active Directory deployments in customer sites consist of two or more domains that represent boundaries between different geographical, managerial, organizational, or administrative layouts. For example, when company "A" acquires company "B", it quickly becomes necessary for preexisting domains to start trusting each other. Alternatively, in some deployments, servers that have a specific role (such as a mail server) can be members of a "resource domain", easing the management burden by combining like roles under one administrative domain.

Communication between disparate domains, especially secure communication that involves authentication and authorization, requires that some stateful knowledge is shared between the peer domains in order for them to trust one another. Some of this knowledge is sensitive, forming the cryptographic basis of trust mechanisms used in protocols such as Kerberos and Netlogon RPC. Other state is public knowledge, such as the NetBIOS name of a peer domain, or which security identifiers are owned by the peer domain. Information like this plays a crucial role when performing name lookups, which are essential for authorization, locating user accounts, or simply displaying information in some type of user interface.

Active Directory stores trust information in trusted domain objects (TDOs) ([MS-ADTS] section and, depending on the kind of trust established, in associated user accounts, interdomain trust accounts, for the trusted domain. There are different types of trusts that exist between Active Directory domains, as described in [MS-ADTS] section TDOs can be managed through the LSAD protocol ([MS-LSAD] section