Volume (computing)

In the context of computer operating systems, volume is the term used to describe a single accessible storage area with a single file system, typically (though not necessarily) resident on a single partition of a hard disk. Similarly, it refers to the logical interface used by an operating system to access data stored on some media using a single instance of a filesystem. "Volume" can be used in place of the term "drive" where it is desirable to indicate that the entity in question is not a physical disk drive, but rather the corporate data stored using a filesystem there. "Logical drive" and "volume" should be considered synonymous, however "volume" and "partition" are not synonymous. In Linux systems, volumes are usually handled by the Logical Volume Manager or the Enterprise Volume Management System and manipulated using [http://www.netadmintools.com/html/8mount.man.html mount(8)] . The term is also used in NT-based versions of Microsoft Windows, where they are handled by the kernel and managed using the Disk Management MMC snap-in.

Differences between volume and partition

A volume is not the same thing as a partition. For example, a floppy disk might be accessible as a volume, even though it does not contain a partition, as floppy disks cannot be partitioned with most modern computer hardware. Also, an OS can recognize a partition without recognizing any volume associated with it, as when the OS cannot interpret the filesystem stored there. This situation occurs, for example, with Windows NT-based OSes and most non-Microsoft OS partitions, such as with the ext3 filesystem commonly used with Linux. Another example occurs in the Intel world with the "Extended Partition". While these are partitions, they cannot contain a filesystem directly. Instead, "logical drives" (aka volumes) must be created within them. This is also the case with NetWare volumes residing inside of a single partition. In short, volumes exist at the logical OS level, and partitions exist at the physical, media specific level. Some times there is a one-to-one correspondence, but it is not guaranteed to be true.

It isn't uncommon to see a volume packed into a single file. Examples include ISO9660 disc images (CD/DVD images, commonly called "ISOs"), and installer volumes for Mac OS X (DMGs). As these volumes are files which reside within another volume, they certainly "aren't" partitions.

Example

This example concerns a Windows XP system with two physical hard disks. The first hard disk has two partitions, the second has only one. The first partition of the first hard disk contains the operating system. Mount points have been left at defaults.

In this example,
* "C:", "D:", and "E:" are volumes.
* Hard Disk 1 and Hard Disk 2 are physical disks.
* Any of these can be called a "drive".

Nomenclature of volumes

Windows-NT based operating systems

It is important to note that Windows NT-based OSes do not have a single root directory. As a result, Windows will assign at least one path to each mounted volume, which will take one of two forms:
* A drive letter, in the form of a single letter followed by a colon, such as "F:"
* A mount-point on an NTFS volume having a drive letter, such as "C:Music"

In these two examples, a file called "Track 1.mp3" stored in the root directory of the mounted volume could be referred to as "F:Track 1.mp3" or "C:MusicTrack 1.mp3" respectively.

In order to assign a mount point for a volume as a path within another volume, the following criteria must be met:
* The volume must be formatted NTFS.
* A folder must exist at the root path. (As of Windows Vista, it can be any subfolder in a volume)
* That folder must be empty.

By default, Windows will assign drive letters to all drives, as follows:
* "A:" and "B:" to floppy disk drives, present or not
* "C:" and subsequent letters, as needed, to:
** Hard disks
** Removable disks, including optical media (e.g. CDs and DVDs)

Because of this convention, the operating system startup drive is most commonly called "C:". This is not always the case.

On Windows XP, mount points may be managed through the Disk Management snap-in for the Microsoft Management Console. This can be most conveniently accessed through "Computer Management" in the "Administrative Tools" section of the Control Panel (Windows).

More than one drive letter can refer to a single volume, as when using the [http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/subst.mspx SUBST] command.

Warning: removing drive letters or mount-points for a drive may break some programs, as some files may not be accessible under the known path. For example, if a program is installed at "D:Program FilesSome Program", it may expect to find its data files at "D:Program FilesSome ProgramData". If the logical disk previously called "D:" has its drive letter changed to "E:", "Some Program" won't be able to find its data at "D:Program FilesSome ProgramData", since the drive letter "D:" no longer represents that volume.

Unix-like operating systems

In Unix-like operating systems, volumes other than the boot volume have a mount-point somewhere within the filesystem, represented by a path. Logically, the directory tree stored on the volume is grafted in at the mountpoint. By convention, mount-points will often be placed in a directory called '/mnt', though '/media' and other terms are sometimes used.

Like in Windows, to use a given path as a mount-point for another volume, an empty directory (sometimes called a folder) must exist there.

Unix-like operating systems use the mount command to manipulate mount points for volumes.

For example, if a CD-ROM drive containing a text file called 'info.txt' was mounted at '/mnt/iso9660', the text file would be accessible at '/mnt/iso9660/info.txt'.

Benefits of keeping files within one volume

Speed of data management

Files within a volume can generally be moved to any other place within that volume by manipulating the filesystem, without moving the actual data. However, if a file is to be moved outside the volume, the data itself must be relocated, which is a much more expensive operation.

In order to better visualize this concept, one might consider the example of a large library. If a non-fiction work is originally classified as having the subject "plants", but then has to be moved to the subject "flora", one does not need to refile the book, whose position on the shelf would be static, but rather, one needs only to replace the index card. However, to move the book to another library, adjusting index cards alone is insufficient. The entire book must be moved.

Special functions of advanced filesystems and volumes

Some filesystems, such as Microsoft's NTFS filesystem, and ext3, allow multiple pseudonyms (known as "hard links") to be created for a single file within the same volume. Hard links allow a file to be referenced by two separate filenames, without its data being stored in two places on the disk (and thereby consuming twice as much space). Hard links cannot be created for files between volumes; this is comparable to moving the file in the library example above. To return to the library analogy, this is like filing two index cards for the same book: one could file the above book under both 'flora' and 'plants'. In general, deleting one hard link does not immediately effect other hard links, while deleting the final hard link for a file frees the disk space occupied by that file. However, modifying the data of the file referred to by one hard link will impact all other hard links as well. In the library, this is comparable to writing in the book. Hard links should not be confused with aliases (Mac OS), shortcuts (Windows), or soft links (Unix and variants), which can refer to files on another volume in some cases. FAT filesystems, such as FAT32, do not support hard or soft links as such, although the Windows operating system supports 'links', which are somewhat less capable.

External links

* MSDN's article on [http://msdn.microsoft.com/en-us/library/aa365006.aspx Hard Links and Junctions]