The Windows Registry is a hierarchical database that stores low-level settings for the Microsoft Windows operating system and for applications that opt to use the registry. The kernel, device drivers, services, Security Accounts Manager, and user interfaces can all use the registry. The registry also allows access to counters for profiling system performance.
In other words, the registry or Windows Registry contains information, settings, options, and other values for programs and hardware installed on all versions of Microsoft Windows operating systems. For example, when a program is installed, a new subkey containing settings such as a program's location, its version, and how to start the program, are all added to the Windows Registry.
When introduced with Windows 3.1, the Windows Registry primarily stored configuration information for COM-based components. Windows 95 and Windows NT extended its use to rationalize and centralize the information in the profusion of INI files, which held the configurations for individual programs, and were stored at various locations. It is not a requirement for Windows applications to use the Windows Registry. For example, .NET Framework applications use XML files for configuration, while portable applications usually keep their configuration files with their executables.
Rationale Prior to the Windows Registry, .INI files stored each program's settings as a text file or binary file, often located in a shared location that did not provide user-specific settings in a multi-user scenario. By contrast, the Windows Registry stores all application settings in one logical repository (but also in a number of discrete files) and in a standardized form. According to Microsoft, this offers several advantages over .INI files. Since file parsing is done much more efficiently with a binary format, it may be read from or written to more quickly than a text INI file. Furthermore, strongly typed data can be stored in the registry, as opposed to the text information stored in .INI files. This is a benefit when editing keys manually using , the built-in Windows Registry Editor. Because user-based registry settings are loaded from a user-specific path rather than from a read-only system location, the registry allows multiple users to share the same machine, and also allows programs to work for less privileged users. Backup and restoration is also simplified as the registry can be accessed over a network connection for remote management/support, including from scripts, using the standard set of APIs, as long as the Remote Registry service is running and firewall rules permit this.
Because the registry is a database, it offers improved system integrity with features such as atomic updates. If two processes attempt to update the same registry value at the same time, one process's change will precede the other's and the overall consistency of the data will be maintained. Where changes are made to .INI files, such race conditions can result in inconsistent data that does not match either attempted update. Windows Vista and later operating systems provide transactional updates to the registry by means of the Kernel Transaction Manager, extending the atomicity guarantees across multiple key or value changes with traditional commit–abort semantics. (Note however that NTFS provides such support for the file system as well, so the same guarantees could, in theory, be obtained with traditional configuration files.)
Structure Keys and values The registry contains two basic elements: keys and values. Registry keys are container objects similar to folders. Registry values are non-container objects similar to files. Keys may contain values and subkeys. Keys are referenced with a syntax similar to Windows' path names, using backslashes to indicate levels of hierarchy. Keys must have a case insensitive name without backslashes.
The hierarchy of registry keys can only be accessed from a known root key handle (which is anonymous but whose effective value is a constant numeric handle) that is mapped to the content of a registry key preloaded by the kernel from a stored "hive", or to the content of a subkey within another root key, or mapped to a registered service or DLL that provides access to its contained subkeys and values.
For example, refers to the subkey "Windows" of the subkey "Microsoft" of the subkey "Software" of the root key.
There are seven predefined root keys, traditionally named according to their constant handles defined in the Win32 API, or by synonymous abbreviations (depending on applications): or or or or or (only in Windows NT, but invisible in the Windows Registry Editor) (only in Windows 9x, and visible in the Windows Registry Editor)
Like other files and services in Windows, all registry keys may be restricted by access control lists (ACLs), depending on user privileges, or on security tokens acquired by applications, or on system security policies enforced by the system (these restrictions may be predefined by the system itself, and configured by local system administrators or by domain administrators). Different users, programs, services or remote systems may only see some parts of the hierarchy or distinct hierarchies from the same root keys.
Registry values are name/data pairs stored within keys. Registry values are referenced separately from registry keys. Each registry value stored in a registry key has a unique name whose letter case is not significant. The Windows API functions that query and manipulate registry values take value names separately from the key path or handle that identifies the parent key. Registry values may contain backslashes in their names, but doing so makes them difficult to distinguish from their key paths when using some legacy Windows Registry API functions (whose usage is deprecated in Win32).
The terminology is somewhat misleading, as each registry key is similar to an associative array, where standard terminology would refer to the name part of each registry value as a "key". The terms are a holdout from the 16-bit registry in Windows 3, in which registry keys could not contain arbitrary name/data pairs, but rather contained only one unnamed value (which had to be a string). In this sense, the Windows 3 registry was like a single associative array, in which the keys (in the sense of both 'registry key' and 'associative array key') formed a hierarchy, and the registry values were all strings. When the 32-bit registry was created, so was the additional capability of creating multiple named values per key, and the meanings of the names were somewhat distorted.Raymond Chen, "Why do registry keys have a default value?" For compatibility with the previous behavior, each registry key may have a "default" value, whose name is the empty string.
Each value can store arbitrary data with variable length and encoding, but which is associated with a symbolic type (defined as a numeric constant) defining how to parse this data. The standard types are:
+ List of standard registry value types Type ID Symbolic type name Meaning and encoding of the data stored in the registry value 0 No type (the stored value, if any) 1 A string value, normally stored and exposed in UTF-16LE (when using the Unicode version of Win32 API functions), usually terminated by a NUL character 2 An "expandable" string value that can contain environment variables, normally stored and exposed in UTF-16LE, usually terminated by a NUL character 3 Binary data (any arbitrary data) 4 A DWORD value, a 32-bit unsigned integer (numbers between 0 and 4,294,967,295 ) (little-endian) 5 A DWORD value, a 32-bit unsigned integer (numbers between 0 and 4,294,967,295 ) (big-endian) 6 A symbolic link (UNICODE) to another registry key, specifying a root key and the path to the target key 7 A multi-string value, which is an ordered list of non-empty strings, normally stored and exposed in Unicode, each one terminated by a null character, the list being normally terminated by a second null character. 8 A resource list (used by the Plug-n-Play hardware enumeration and configuration) 9 A resource descriptor (used by the Plug-n-Play hardware enumeration and configuration) 10 A resource requirements list (used by the Plug-n-Play hardware enumeration and configuration) 11 A QWORD value, a 64-bit integer (either big- or little-endian, or unspecified) (introduced in Windows 2000)
Root keys The keys at the root level of the hierarchical database are generally named by their Windows API definitions, which all begin with They are frequently abbreviated to a three- or four-letter short name starting with "HK" (e.g. and ). Technically, they are predefined handles (with known constant values) to specific keys that are either maintained in memory, or stored in hive files stored in the local filesystem and loaded by the system kernel at boot time and then shared (with various access rights) between all processes running on the local system, or loaded and mapped in all processes started in a user session when the user logs on the system.
The (local machine-specific configuration data) and (user-specific configuration data) nodes have a similar structure to each other; user applications typically look up their settings by first checking for them in , and if the setting is not found, look instead in the same location under the key. However, the converse may apply for administrator-enforced policy settings where HKLM may take precedence over . The Windows Logo Program has specific requirements for where different types of user data may be stored, and that the concept of least privilege be followed so that administrator-level access is not required to use an application.
HKEY_LOCAL_MACHINE (HKLM) Abbreviated , stores settings that are specific to the local computer.
The key located by is actually not stored on disk, but maintained in memory by the system kernel in order to map all the other subkeys. Applications cannot create any additional subkeys. On Windows NT, this key contains four subkeys—, , , and —that are loaded at boot time within their respective files located in the folder. A fifth subkey, , is volatile and is created dynamically, and as such is not stored in a file (it exposes a view of all the currently detected Plug-and-Play devices). On Windows Vista and above, a sixth and seventh subkey— and —are mapped in memory by the kernel on-demand and loaded from or from boot configuration data, on the system partition.
The key usually appears as empty for most users (unless they are granted access by administrators of the local system or administrators of domains managing the local system). It is used to reference all "Security Accounts Manager" (SAM) databases for all domains into which the local system has been administratively authorized or configured (including the local domain of the running system, whose SAM database is stored in a subkey also named "SAM": other subkeys will be created as needed, one for each supplementary domain). Each SAM database contains all built in accounts (mostly group aliases) and configured accounts (users, groups and their aliases, including guest accounts and administrator accounts) created and configured on the respective domain, for each account in that domain, it notably contains the user name which can be used to log on that domain, the internal unique user identifier in the domain, a cryptographic hash of each user's password for each enabled authentication protocol, the location of storage of their user registry hive, various status flags (for example if the account can be enumerated and be visible in the logon prompt screen), and the list of domains (including the local domain) into which the account was configured.