ReFS is a huge improvement over the current industry standard for Windows machines, NTFS.  It is an acronym for Resilient File System and was originally codenamed “Protogon”. The intended use case is for file servers and most notably offers two major improvements:

  • Improved reliability
  • Built-in resilience

The increased resilience is found in an allocation-on-write update process for metadata that allocates new areas for every update transaction and uses large IO batches.  All ReFS metadata has built-in 64-bit checksums.  This built-in resiliency has deprecated the need to run error-checking tools such as CHKDSK when using ReFS.  The repairs can be made while the volume is online and no longer needs to take the volume offline for maintenance!

More information can be found in a MSDN Blog: