Home > Understanding the Operating System and File System > FAT

FAT

FAT by George Hernandez
Modified 2003/05/27


File Allocation Tables are used by the OS to point to and find all the parts of all the files on your hard drive. Here are three Microsoft FATs in historical order:


FAT (Aka Virtual FAT or FAT 16)


Can be used by DOS, Win 3.x/9x/NT/2000 and OS/2.


Names follow the 8.3 naming convention. It must be an ASCII character except for spaces and the following: . " / \ [ ] : ; | = ,. All characters are effectively uppercase.


Works well with volumes less than 500 MB, becomes impractical when a volume reaches 1 GB, and has a maximum file and volume size of 2^32 B (4 GB) in NT and 2 GB in DOS.


Has 16 b cluster addressing and is thus limited to 2^16 or 65,535 clusters (entries) per volume.


Sectors are 512 B.


Logical drives should be limited to 2.1 GB.


Can enable share-level security on the directory level (access to any user via password).


NTFS (NT File System)


Can be used by Win NT/2000.


Names can be up to 255 characters including extensions. It must be any Unicode character except for the following: ? " / \ < > * | :. Although it is only practical to use a Windows ANSI character. Case is preserved but effectively irrelevant.


Works well with volumes above 400 MB and has a theoretical maximum file and volume size of 2^64 B (18 EB).


Has 64 b cluster addressing and is thus limited to 2^64 or roughly 1.8E19 clusters (entries) per volume.


Sectors can vary in size.


Allows permissions for individuals and groups to access individual files and folders.


Enables individual file and folder compression, as well as usage of compressed items without having to decompress.


FAT 32


Can be used by Win 98/2000.


Works well up to 1 TB.


Has 32 b cluster addressing.


Logical drives should be limited to 2 TB.


Here is a recommended layout for utilizing logical drives:


  • C: for operating system and utilities.
  • D: for applications.
  • E: for data.
  • F: for anything that needs a FAT drive.

This way if a failure occurs, it will usually occur on drives C: or D:, leaving data on drive E: intact. The FAT drive F: can be used by most operating systems.