Comparison of file systems 1 Comparison of file systems The following tables compare general and technical information for a number of file systems. General information File system Creator Year introduced Original operating system DECtape DEC 1964 PDP-6 Monitor Level-D DEC 1968 TOPS-10 George 2 ICT (later ICL) 1968 George 2 V6FS Bell Labs 1972 Version 6 Unix ODS-1 DEC 1972 RSX-11 RT-11 file system DEC 1973 RT-11 DOS (GEC) GEC 1973 Core Operating System CP/M file system Gary Kildall 1974 CP/M OS4000 GEC 1977 OS4000 FAT (8-bit) Marc McDonald, Microsoft 1977 Microsoft Disk BASIC DOS 3.x Apple Computer 1978 Apple DOS Pascal Apple Computer 1978 Apple Pascal CBM DOS Commodore 1978 Microsoft BASIC (for CBM PET) V7FS Bell Labs 1979 Version 7 Unix ODS-2 DEC 1979 OpenVMS FAT12 Tim Paterson 1980 QDOS, 86-DOS AFS Carnegie Mellon University 1982 Multiplatform MultoOS DFS Acorn Computers Ltd 1982 Acorn BBC Micro MOS ADFS Acorn Computers Ltd 1983 Acorn Electron (later Arthur RISC OS) FFS Kirk McKusick 1983 4.2BSD ProDOS Apple Computer 1983 ProDOS 8 MFS Apple Computer 1984 Mac OS FAT16 Microsoft 1984 MS-DOS 3.0 Elektronika BK tape format NPO "Scientific centre" (now Sitronics) 1985 Vilnius Basic, BK monitor program HFS Apple Computer 1985 Mac OS Amiga OFS[1] Metacomco for Commodore 1985 Amiga OS High Sierra Ecma International 1985 MS-DOS, Mac OS NWFS Novell 1985 NetWare 286 FAT16B Compaq 1987 Compaq MS-DOS 3.31, DR DOS 3.31 MINIX V1 FS Andrew S. Tanenbaum 1987 MINIX 1.0 Amiga FFS Commodore 1988 Amiga OS 1.3
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Comparison of file systems 1
Comparison of file systemsThe following tables compare general and technical information for a number of file systems.
General information
File system Creator Yearintroduced
Original operating system
DECtape DEC 1964 PDP-6 Monitor
Level-D DEC 1968 TOPS-10
George 2 ICT (later ICL) 1968 George 2
V6FS Bell Labs 1972 Version 6 Unix
ODS-1 DEC 1972 RSX-11
RT-11 file system DEC 1973 RT-11
DOS (GEC) GEC 1973 Core Operating System
CP/M file system Gary Kildall 1974 CP/M
OS4000 GEC 1977 OS4000
FAT (8-bit) Marc McDonald, Microsoft 1977 Microsoft Disk BASIC
DOS 3.x Apple Computer 1978 Apple DOS
Pascal Apple Computer 1978 Apple Pascal
CBM DOS Commodore 1978 Microsoft BASIC (for CBM PET)
V7FS Bell Labs 1979 Version 7 Unix
ODS-2 DEC 1979 OpenVMS
FAT12 Tim Paterson 1980 QDOS, 86-DOS
AFS Carnegie Mellon University 1982 Multiplatform MultoOS
DFS Acorn Computers Ltd 1982 Acorn BBC Micro MOS
ADFS Acorn Computers Ltd 1983 Acorn Electron (later Arthur RISC OS)
FFS Kirk McKusick 1983 4.2BSD
ProDOS Apple Computer 1983 ProDOS 8
MFS Apple Computer 1984 Mac OS
FAT16 Microsoft 1984 MS-DOS 3.0
Elektronika BK tapeformat
NPO "Scientific centre" (nowSitronics)
1985 Vilnius Basic, BK monitor program
HFS Apple Computer 1985 Mac OS
Amiga OFS[1] Metacomco for Commodore 1985 Amiga OS
High Sierra Ecma International 1985 MS-DOS, Mac OS
NWFS Novell 1985 NetWare 286
FAT16B Compaq 1987 Compaq MS-DOS 3.31, DR DOS 3.31
Notes[1] http:/ / en. wikipedia. org/ wiki/ Comparison_of_file_systems#endnote_54[2] IBM introduced JFS with the initial release of AIX Version 3.1 in 1990. This file system now called JFS1. The new JFS, ported from OS/2 to
AIX and Linux, was first shipped in OS/2 Warp Server for e-Business in 1999. It was released as JFS2 on AIX 5L.[4] Microsoft first introduced FAT32 in Windows 95 OSR2 (OEM Service Release 2) and then later in Windows 98. NT-based Windows did not
have any support for FAT32 up to Windows NT4; Windows 2000 was the first NT-based Windows OS that received the ability to work withit.
[5] Specifications for the Reliance file systems are available here (http:/ / www. datalight. com/ products/ filesystems/reliance-family-specifications).
[6] These are the restrictions imposed by the on-disk directory entry structures themselves. Particular Installable File System drivers may placerestrictions of their own on file and directory names; and particular and operating systems may also place restrictions of their own, across allfilesystems. MS-DOS, Microsoft Windows, and OS/2 disallow the characters UNIQ-nowiki-0-a3de2c3c417b2102-QINU in file and directorynames across all filesystems. Unix-like systems disallow the characters UNIQ-nowiki-1-a3de2c3c417b2102-QINU in file and directory namesacross all filesystems.
[7] For filesystems that have variable allocation unit (block/cluster) sizes, a range of size are given, indicating the maximum volume sizes for theminimum and the maximum possible allocation unit sizes of the filesystem (e.g. 512 bytes and 128 kB for FAT — which is the cluster sizerange allowed by the on-disk data structures, although some Installable File System drivers and operating systems do not support cluster sizeslarger than 32 kB).
[8][8] While the on-disk filesystem structure uses a 4-byte file length, which allows files up to 4G, the usual disk access APIs use the top three bitsof the sector number to specify the drive number, effectively limiting the maximum file size to 512M.
[9][9] While the on-disk filesystem structure uses a 3-byte sector number, which allows access to 4G of disk space, the usual disk access APIs usethe top three bits of the sector number to specify the drive number, effectively limiting the maximum disk size to 512M.
[10][10] The CP/M filesystem itself does have limitations in regard to the allowed filename characters to be used, but officially the followingcharacters are not allowed: UNIQ-nowiki-2-a3de2c3c417b2102-QINU . CCP reserves the following characters for special purposes:UNIQ-nowiki-3-a3de2c3c417b2102-QINU , PIP additionally reserves: UNIQ-nowiki-4-a3de2c3c417b2102-QINU .
[12] Depends on whether the FAT12, FAT16, and FAT32 implementation has support for long filenames (LFNs). Where it does not, as in OS/2,MS-DOS, Windows 95, Windows 98 in DOS-only mode and the Linux "msdos" driver, file names are limited to 8.3 format of 8-bit characters(space padded in both the basename and extension parts) and may not contain NUL (end-of-directory marker) or character 5 (replacement forcharacter 229 which itself is used as deleted-file marker). Short names also do not normally contain lowercase letters. Also note that a fewspecial names (CON, NUL, LPT1) should be avoided, as some operating systems (notably DOS and windows) effectively reserve them.
[13][13] In these filesystems the directory entries named "." and ".." have special status. Directory entries with these names are not prohibited, andindeed exist as normal directory entries in the on-disk data structures. However, they are mandatory directory entries, with mandatory values,that are automatically created in each directory when it is created; and directories without them are considered corrupt.
[14] The on-disk structures have no inherent limit. Particular Installable File System drivers and operating systems may impose limits of theirown, however. MS-DOS/PC DOS do not support full pathnames longer than 66 bytes for FAT12, FAT16 and FAT32 volumes. This limitexists because these operating systems were designed around a fixed-length internal data structure named Current Directory Structure, whichholds the absolute paths of the current working directories of all volumes. The FAT12/FAT16 file system implementation under ConcurrentDOS and DR DOS 3.31 to 6.0 (prior to 1992 updates) did not impose any such limits on the directory depth due to their internal representationof current working directories as dynamically updated chain of double-linked relative directories. The introduction of a DOS-like CDS(instead of only an emulation thereof) for compatibility purposes with BDOS 7.0 in 1992 imposed the same length limits on PalmDOS, DRDOS 6.0 (since 1992 update), Novell DOS, OpenDOS, etc. as known from MS-DOS/PC DOS. Windows NT does not support full pathnameslonger than 32,767 bytes for NTFS. Most Windows programs will fail when full path exceeds 255 characters (including Explorer andCMD.EXE). Linux has a pathname limit of 4,096.
[15] See manual http:/ / wwwlehre. dhbw-stuttgart. de/ ~helbig/ os/ v6/ doc/ V/ fs. html[16] The actual maximum was 1,082,201,088 bytes, with 10 direct blocks, 1 singly indirect block, 1 doubly indirect block, and 1 triply indirect
block. The 4.0BSD and 4.1BSD versions, and the System V version, used 1,024-byte blocks rather than 512-byte blocks, making themaximum 4,311,812,608 bytes or approximately 4 GB.
[17] Table "Limits" states a maximum of 255 Unicode characters for the filename (http:/ / msdn. microsoft. com/ en-us/ library/ ee681827(VS.85). aspx)
[19] While FAT32 partitions this large work fine once created, some software won't allow creation of FAT32 partitions larger than 32 GB. Thisincludes, notoriously, the Windows XP installation program and the Disk Management console in Windows 2000, XP, 2003 and Vista. UseFDISK from a Windows ME Emergency Boot Disk to avoid. (http:/ / support. microsoft. com/ kb/ 314463)
[20] As Mac OS X is a Unix-like system, which supports UNIQ-nowiki-5-a3de2c3c417b2102-QINU in file names, and which usesUNIQ-nowiki-6-a3de2c3c417b2102-QINU as a pathname component separator, SPACE $ & % @ \ ^ : . # * " ¦ in file names is represented ondisk in HFS and HFS+ as UNIQ-nowiki-8-a3de2c3c417b2102-QINU .
[21] The "." and ".." directory entries in HPFS that are seen by applications programs are a partial fiction created by the Installable File Systemdrivers. The on-disk data structure for a directory does not contain entries by those names, but instead contains a special "start" entry. Whilston-disk directory entries by those names are not physically prohibited, they cannot be created in normal operation, and a directory containingsuch entries is corrupt.
[22] This is the limit of the on-disk structures. The HPFS Installable File System driver for OS/2 uses the top 5 bits of the volume sector numberfor its own use, limiting the volume size that it can handle to 64 GB.
[23] NTFS allows files to have multiple names, in separate namespaces: Win32, DOS, Win32&DOS, and Posix. Windows APIs create files withWin32 "long" names (1–255 characters), sometimes with an additional "short"/"alias" DOS name in the "8.3" format (12 characters).
[24] NB: This article includes discussion of the NT & Win32 namespaces used by Windows APIs; these are distinct from the NTFS filenamenamespaces.
[25] In the Win32 namespace, any UTF-16 code unit (case insensitive) except NUL and UNIQ-nowiki-9-a3de2c3c417b2102-QINU is allowed;in the Posix namespace, any UTF-16 code unit (case sensitive) except NUL and / is allowed; in the DOS namespace, any character in theU+0021–U+007E range except SPACE < > . , ; : = ? * [ ] % | ( ) / \ is allowed. Windows APIs require Win32 namespace compatibility, whichprevents access to folders & files having only Posix names containing Win32-incompatible characters.
[26] This is the limit of the on-disk structures. The NTFS driver for Windows NT limits the volume size that it can handle to 256 TB and the filesize to 16 TB respectively.
[28][28] The Mac OS provides two sets of functions to retrieve file names from an HFS Plus volume, one of them returning the full Unicode names,the other shortened names fitting in the older 31 byte limit to accommodate older applications.
[29] HFS Plus mandates support for an escape sequence to allow arbitrary Unicode. Users of older software might see the escape sequencesinstead of the desired characters.
[31] ext4 1.42 (http:/ / e2fsprogs. sourceforge. net/ e2fsprogs-release. html#1. 42) "This release of e2fsprogs has support for file systems > 16TB"
[32][32] Depends on kernel version and arch. For 2.4 kernels the max is 2 TB. For 32-bit 2.6 kernels it is 16 TB. For 64-bit 2.6 kernels it is 8 EB.[33] ReiserFS has a theoretical maximum file size of 1 EB, but "page cache limits this to 8TB on architectures with 32 bit int" (http:/ / www.
namesys. com/ faq. html#reiserfsspecs)[34][34] QFS allows files to exceed the size of disk when used with its integrated HSM, as only part of the file need reside on disk at any one time.[35][35] Varies wildly according to block size and fragmentation of block allocation groups.[36][36] NSS allows files to have multiple names, in separate namespaces.[37] Some namespaces had lower name length limits. "LONG" had an 80-byte limit, "NWFS" 80 bytes, "NFS" 40 bytes and "DOS" imposed 8.3
filename.[38][38] Maximum combined filename/filetype length is 236 bytes; each component has an individual maximum length of 255 bytes.[39][39] Maximum pathname length is 4,096 bytes, but quoted limits on individual components add up to 1,664 bytes.[40][40] This restriction might be lifted in newer versions.[41] 232 × block size[42][42] Maximum file size on a VMFS volume depends on the block size for that VMFS volume. The figures here are obtained by using the
maximum block size.[43][43] ISO 9660#Restrictions[44] Through the use of multi-extents, a file can consist of multiple segments, each up to 4 GB in size. See ISO 9660#The 2/4 GB file size limit[45][45] Assuming the typical 2048 Byte sector size. The volume size is specified as a 32-bit value identifying the number of sectors on the volume.[46] Joliet Specification (http:/ / bmrc. berkeley. edu/ people/ chaffee/ jolspec. html)[47] https:/ / raw. github. com/ danrl/ lanyfs-docs/ master/ lanyfs-1. 4. txt[49] Note that the filename can be much longer XFS#Extended_attributes[50] XFS has a limitation under Linux 2.4 of 64 TB file size, but Linux 2.4 only supports a maximum block size of 2 TB. This limitation is not
present under IRIX.[51][51] Implemented in later versions as an extension[52] Concurrent DOS, FlexOS, Multiuser DOS, REAL/32, PalmDOS, Novell DOS, OpenDOS, and DR-DOS can store file owner information in
reserved fields of directory entries on FAT12 and FAT16 volumes, if the optional multi-user security module is loaded. If loaded, mostexternal commands invoke support for special /U:owner/group command line options to deal with this extra information.
[53] Concurrent DOS, FlexOS, Multiuser DOS, REAL/32, DR DOS, PalmDOS, Novell DOS, OpenDOS, and DR-DOS can storeread/write/delete/execute access permissions and file/directory passwords in reserved fields of directory entries on FAT12 and FAT16volumes. This is an integral part of the design, therefore passwords can be appended to file or directory names with semicolon (for example:dirname;dirpwd\filename;filepwd), the PASSWORD command can be used to control permissions and some commands support a special/P:pwd option to deal with this feature.
[54][54] File creation and file access timestamps are supported only by DOS 7.0 and higher, and typically only when explicitly enabled.[55][55] Some FAT implementations, such as in Linux, show file modification timestamp (mtime) in the metadata change timestamp (ctime) field.
This timestamp is however, not updated on file metadata change.[56] Particular Installable File System drivers and operating systems may not support extended attributes on FAT12 and FAT16. The OS/2 and
Windows NT filesystem drivers for FAT12 and FAT16 support extended attributes (using a "EA DATA. SF" pseudo-file to reserve theclusters allocated to them). Other filesystem drivers for other operating systems do not.
[57] The f-node contains a field for a user identifier. This is not used except by OS/2 Warp Server, however.[58] NTFS access control lists can express any access policy possible using simple POSIX file permissions (and far more), but use of a
POSIX-like interface is not supported without an add-on such as Services for UNIX or Cygwin.[59] As of Vista, NTFS has support for Mandatory Labels, which are used to enforce Mandatory Integrity Control. See (http:/ / msdn2. microsoft.
com/ en-us/ library/ bb648648. aspx)[61][61] Access-control lists and MAC labels are layered on top of extended attributes.[62][62] Some operating systems implemented extended attributes as a layer over UFS1 with a parallel backing file (e.g., FreeBSD 4.x).[63] Some Installable File System drivers and operating systems may not support extended attributes, access control lists or security labels on
these filesystems. Linux kernels prior to 2.6.x may either be missing support for these altogether or require a patch.[64] ext4 has group descriptor, journal and, starting from Linux kernel 3.5, metadata checksumming[65][65] Creation time is stored in the backing ext4 filesystem, but is not yet sent to clients.[66][66] Lustre has checksums for data over the network, but depends on backing filesystem and hardware for checksums of persistent data[67][67] Not available with ext3/4, but will be available with ZFS OST/MDT backing filesystems.[68][68] ocfs2 computes and validates checksums of metadata objects like inodes and directories. It also stores an error correction code capable to
fixing single-bite errors.[69][69] CRCs are employed for certain types of metadata.[70] The local time, timezone/UTC offset, and date are derived from the time settings of the reference/single timesync source in the NDS tree.[71][71] Novell calls this feature "multiple data streams". Published specifications say that NWFS allows for 16 attributes and 10 data streams, and
NSS allows for unlimited quantities of both.[72] Some file and directory metadata is stored on the NetWare server irrespective of whether Directory Services is installed or not, like
date/time of creation, file size, purge status, etc; and some file and directory metadata is stored in NDS/eDirectory, like file/objectpermissions, ownership, etc.
[73][73] Record Management Services (RMS) attributes include record type and size, among many others.[74] File permission in 9P are a variation of the traditional Unix permissions with some minor changes, e.g. the suid bit is replaced by a new
'exclusive access' bit.[75][75] MAC/Sensitivity labels are per filesystem. A label per file are not out of the question as a future compatible change but aren't part of any
available version of ZFS.[76][76] Solaris "extended attributes" are really full-blown alternate data streams, in both the Solaris UFS and ZFS. ZFS also has "system attributes"
used for storing MS-DOS/NTFS compatible attributes for use by CIFS; as well as some attributes ported from FreeBSD[77][77] Time the file was recorded on the volume always available; "File Creation Date and Time" available only if the file has an Extended
Attribute block.[78][78] Not applicable to file systems on a read-only medium.[79][79] Available only if the file has an Extended Attribute block.[80][80] Symlinks only visible to NFS clients. References and Off-Disk Pointers (ODPs) provide local equivalent.[81] System V Release 4, and some other Unix systems, retrofitted symbolic links to their versions of the Version 7 Unix file system, although
the original version didn't support them.[82][82] Context based symlinks were supported in GFS, GFS2 only supports standard symlinks since the bind mount feature of the Linux VFS has
made context based symlinks obsolete[83][83] Optional journaling of data[84] As of Windows Vista, NTFS fully supports soft links. See this Microsoft article on Vista kernel improvements (http:/ / www. microsoft.
com/ technet/ technetmag/ issues/ 2007/ 02/ VistaKernel/ default. aspx). NTFS 5.0 (Windows 2000) and higher can create junctions (http:/ /support. microsoft. com/ kb/ 205524), which allow any valid local directory (but not individual files) ("target" of junction) to be mapped to anNTFS version thereof ("source" = location of junction). The source directory must lie on an NTFS 5+ partition, but the target directory can lieon any valid local partition and needn't be NTFS. Junctions are implemented through reparse points, which allow the normal process offilename resolution to be extended in a flexible manner.
[85][85] NTFS stores everything, even the file data, as meta-data, so its log is closer to block journaling.[86] While NTFS itself supports case sensitivity, the Win32 environment subsystem cannot create files whose names differ only by case for
compatibility reasons. When a file is opened for writing, if there is any existing file whose name is a case-insensitive match for the new file,the existing file is truncated and opened for writing instead of a new file with a different name being created. Other subsystems like e. g.Services for Unix, that operate directly above the kernel and not on top of Win32 can have case-sensitivity.
[87] NTFS does not internally support snapshots, but in conjunction with the Volume Shadow Copy Service can maintain persistent blockdifferential volume snapshots.
[88][88] Mac OS System 7 introduced the 'alias', analogous to the POSIX symbolic link but with some notable differences. Not only could they crossfile systems but they could point to entirely different file servers, and recorded enough information to allow the remote file system to bemounted on demand. It had its own API that application software had to use to gain their benefits-- this is the opposite approach from POSIXwhich introduced specific APIs to avoid the symbolic link nature of the link. The Finder displayed their file names in an italic font (at least inRoman scripts), but otherwise they behaved identically to their referent.
[90][90] Metadata-only journaling was introduced in the Mac OS 10.2.2 HFS Plus driver; journaling is enabled by default on Mac OS 10.3 and later.
[91] Although often believed to be case sensitive, HFS Plus normally is not. The typical default installation is case-preserving only. From MacOS 10.3 on the command newfs_hfs -s (http:/ / developer. apple. com/ documentation/ Darwin/ Reference/ ManPages/ man8/ newfs_hfs. 8.html) will create a case-sensitive new file system. HFS Plus version 5 optionally supports case-sensitivity. However, since case-sensitivity isfundamentally different from case-insensitivity, a new signature was required so existing HFS Plus utilities would not see case-sensitivity as afile system error that needed to be corrected. Since the new signature is 'HX', it is often believed this is a new filesystem instead of a simply anupgraded version of HFS Plus. See Apple's File System Comparisons (http:/ / developer. apple. com/ documentation/ MacOSX/ Conceptual/BPFileSystem/ Articles/ Comparisons. html) (which hasn't been updated to discuss HFSX) and Technical Note TN1150: HFS Plus VolumeFormat (http:/ / developer. apple. com/ technotes/ tn/ tn1150. html) (which provides a very technical overview of HFS Plus and HFSX).
[92] Mac OS Tiger (10.4) and late versions of Panther (10.3) provide file change logging (it's a feature of the file system software, not of thevolume format, actually). See fslogger (http:/ / www. kernelthread. com/ software/ fslogger/ ).
[93][93] As of OS X 10.7, HFS+ supports full volume file encryption known as Filevault 2.[94] Since Mac OS X Snow Leopard, online resizing is supported.[95] "Write Ahead Physical Block Logging" in NetBSD, provides metadata journaling and consistency as an alternative to softdep.[97] "Soft dependencies" (softdep) in NetBSD, called "soft updates" in FreeBSD provide meta-data consistency at all times without double writes
(journaling).[98][98] Block level journals can be added by using gjournal module in FreeBSD.[100] UDF, LFS, and NILFS are log-structured file systems and behave as if the entire file system were a journal.[101][101] Linux kernel versions 2.6.12 and newer.[102][102] Offline growing/shrinking as well as online growing:[103][103] Off by default.[104][104] Can be shrunk online by migrating files off an OST and removing the OST, or offline with ext3/4 backing filesystems by shrinking the
OST filesystem[105][105] Full block journaling for ReiserFS was not added to Linux 2.6.8 for obvious reasons.[106] Reiser4 supports transparent compression and encryption with the cryptcompress plugin which is the default file handler in version 4.1.[107][107] OCFS2 supports creating multiple write-able snapshots of regular files using REFLINK.[108][108] File system implements reliability via atomic transactions.[109][109] Optionally no on IRIX.[110] Particular Installable File System drivers and operating systems may not support case sensitivity for JFS. OS/2 does not, and Linux has a
mount option for disabling case sensitivity.[111] (http:/ / www. linux. com/ archive/ feed/ 32002)[112][112] Case-sensitivity/Preservation depends on client. Windows, DOS, and OS/2 clients don't see/keep case differences, whereas clients
accessing via NFS or AFP may.[113] The file change logs, last entry change timestamps, and other filesystem metadata, are all part of the extensive suite of auditing capabilities
built into NDS/eDirectory called NSure Audit. ( Filesystem Events tracked by NSure (http:/ / www. novell. com/ documentation/ nsureaudit/html/ netware_event_data. htm))
[114][114] Available only in the "NFS" namespace.[115] Limited capability. Volumes can span physical disks (volume segment)[116][116] These are referred to as "aliases".[117][117] VxFS provides an optional feature called "Storage Checkpoints" which allows for advanced file system snapshots.[118] When used with venti.[119][119] ZFS is a transactional filesystem using copy-on-write semantics, guaranteeing an always-consistent on-disk state without the use of a
traditional journal. However, it does also implement an intent log to provide better performance when synchronous writes are requested.[121] Variable block size refers to systems which support different block sizes on a per-file basis. (This is similar to extents but a slightly
different implementational choice.) The current implementation in UFS2 is read-only.[122][122] only for .REL (record structured) files, up to 254 bytes/record[123][123] Btrfs can only inline files smaller than 3916B with its metadata[124] SuperStor in DR DOS 6.0 and PC DOS 6.1, DoubleSpace in MS-DOS 6.0, DriveSpace in MS-DOS 6.22, Windows 95 and Windows 98,
and Stacker in Novell DOS 7, OpenDOS 7.01, DR-DOS 7.02/7.03 and PC DOS 7.0/2000 were data compression schemes for FAT.[125][125] Only for "stuffed" inodes[126][126] Only if formatted with 4kB-sized clusters or smaller[127][127] Other block:fragment size ratios supported; 8:1 is typical and recommended by most implementations.[128][128] Fragments were planned, but never actually implemented on ext2 and ext3.[129] e2compr, a set of patches providing block-based compression for ext2, has been available since 1997, but has never been merged into the
mainline Linux kernel.[130][130] In "extents" mode.[132][132] Each possible size (in sectors) of file tail has a corresponding suballocation block chain in which all the tails of that size are stored. The
overhead of managing suballocation block chains is usually less than the amount of block overhead saved by being able to increase the blocksize but the process is less efficient if there is not much free disk space.
[134][134] When enabled, ZFS's logical-block based compression behaves much like tail-packing for the last block of a file.[135] Files, Databases, and Persistent Storage (http:/ / msdn. microsoft. com/ en-us/ library/ ms899821. aspx). MSDN.[136][136] Via dosFs.[137][137] Native FAT32 support with MS-DOS 7.10 and 8.0. Loadable FAT32 support for any DOS since 3.31 with DRFAT32 redirector driver.
Native FAT32 support since OEM DR-DOS 7.04, bootable FAT32 support since OEM DR-DOS 7.06. Native FAT32 support with OEM PCDOS 7.10.
[138] Tuxera NTFS for Windows CE. See article (http:/ / www. windowsfordevices. com/ c/ a/ News/ Tuxera-NTFS-for-Windows-CE/ ) andannouncement (http:/ / www. tuxera. com/ about-us/ news/ 1915/ ).
[139] Win NT 4.0 HPFS Driver (http:/ / hobbes. nmsu. edu/ h-viewer. php?dir=/ pub/ windows& file=hpfsnt. zip)[140] Ext2Fsd is an open source ext2/ext3/ext4 kernel-level file system driver for Windows systems (NT/2K/XP/VISTA/7, X86/AMD64) that
provides both read/write access to the file system. Currently, does not fully support extents (no size truncating/extending, no file deletion), adefault feature of ext4. (http:/ / www. ext2fsd. com/ )
[141] Ext2 IFS for Windows provides kernel-level read/write access to ext2 and ext3 volumes in Windows NT4, 2000, XP, Vista and Windows2008. Does not support inodes size above 128 bytes and does not support ext4. (http:/ / www. fs-driver. org/ faq. html)
[142] Ext2Read is an explorer-like utility to explore ext2/ext3/ext4 file systems that provides read-only access to the file system. It supportsextents, large inodes, and LVM2 volumes. Ext2Read (http:/ / sourceforge. net/ projects/ ext2read/ )
[143] Fuse-ext2 is a multi OS FUSE module to mount ext2 and ext3 file system devices and/or images with read and write support. (http:/ /fuse-ext2. sourceforge. net/ )
[144] Paragon ExtFS for Mac is a low-level file system driver specially developed to bridge file system incompatibility between Linux and Macby providing full read/write access to the Ext2, Ext3 and Ext4 file systems under Mac OS X. (http:/ / www. paragon-software. com/ home/extfs-mac/ )
[145] Ext2fsx is the first and old implementation of the Ext2 (Linux) filesystem for Mac OS X. (http:/ / sourceforge. net/ projects/ ext2fsx/ )[146] OS/2 ext2 Driver (http:/ / hobbes. nmsu. edu/ h-viewer. php?dir=/ pub/ os2/ system/ drivers/ filesys& file=ext2_240. zip)[147] See Total Commander, which supports accessing ext2, ext3, and ReiserFS from Windows, Windows CE, and Windows Mobile.[148] Native ZFS for Linux (http:/ / zfsonlinux. org/ )[149] ZFS on FUSE (http:/ / www. wizy. org/ wiki/ ZFS_on_FUSE)[150] Mac ZFS (http:/ / code. google. com/ p/ maczfs/ )[151] http:/ / wiki. lustre. org/ index. php/ Windows_Native_Client[152] http:/ / wiki. lustre. org/ index. php?title=Main_Page[153] http:/ / wiki. lustre. org/ index. php/ FAQ_-_OS_Support[154] Using SAM-QFS on Linux Clients (http:/ / wikis. sun. com/ display/ SAMQFSDocs/ Using+ SAM-QFS+ on+ Linux+ Clients)[155] ncpfs (http:/ / freshmeat. net/ projects/ ncpfs/ )[157] (http:/ / www-03. ibm. com/ systems/ software/ gpfs/ )[158] vmfs (http:/ / code. google. com/ p/ vmfs/ )[159] AncientFS (http:/ / osxbook. com/ software/ ancientfs/ )[160] VMS2Linux (http:/ / www. vms2linux. de/ )[161] logfs (http:/ / logfs. sourceforge. net/ )
External links• Linux kernel file systems (http:/ / howto. wikia. com/ wiki/ Howto_configure_the_Linux_kernel/ fs) via
Wikia:en.howto:Wikihowto• A speed comparison of filesystems on Linux 2.4.5 (http:/ / web. archive. org/ web/ 20040407211142/ aurora.