PAR (parity) files are created from a source file or files (recovery set), and can be used to restore the recovery set if any file becomes corrupted.

PAR

PAR files were widespreadly used for Usenet newsgroup multi-part archive files. When posting a file a user could also post PAR files which would allow downloaders to recover any part of the archive by downloading an equal number of PAR files. I.e.: if two parts of the archive were missing or corrupt the downloader would just need to download any two PAR files and they could successfully recreate the archive.

PAR files are named as follows:

.PAR
.P01
.P02
.P03
etc.

The .PAR file is simply a small (~1KiB) index file for the set, and does not contain any parity data.

PAR2

PAR2 succeeded PAR; its main benefit being that there is no relationship between the size of the source files and the size of the PAR2 files, and that undamaged blocks in a corrupt PAR2 file can still be used to recover the recovery set. The QuickPar website has a good desscription of the main differences between the two formats.

PAR2 files are named as follows:

filename.volnn+mm.PAR2

nn is the total number of blocks of all preceding files
mm is the number of blocks in this file, and therefore the number of blocks it can repair

How Does It Work?

The redundant data in the PAR files is computed using the Reed-Solomon algorithm that allows for recovery of any 'X' real data-blocks for 'X' parity data-blocks present. (data-blocks referring to files or much smaller virtual slices of files).

For more information read the full PAR 2.0 specification.

Useful Links

Parchive: Parity Archive Tool : Home of PAR
QuickPar : A superb PAR/PAR2 GUI
dvdisaster : Although not exactly PAR, dvdisaster uses the Reed-Solomon algorithm to create a correction file for a DVD image