Redundant Array of Independent FilesystemsUsing the fan-out infrastructure we have, we are developing a file system
that has the same redundancy characteristics of RAID, but at the VFS level.
We call this Redundant Arrays of Independent Filesystems (RAIF).
RAIF support includes striping (RAIF0), mirroring (RAIF1), parity (RAIF5 and
RAIF6), and other modes. RAIF allows redundancy and performance increases
on many types of file systems, including NFS. The RAIF logic is at the VFS
layer, so file-level knowledge is used to decide on the right level of
redundancy that best matches the value of the file in question. Moreover,
recovery from failures can be done on a per file basis instead of the whole
disk drive; this means that if a recovery failed mid-way, it can be resumed
without having to restart the entire disk device's recovery.
If you find a problem, please describe it in detail and submit a bug report.
Conference and Workshop Papers:
||Title (click for html version)
||Storage Virtualization with a Stackable File System
||PDF BibTeX ||Stony Brook U. CS TechReport FSL-05-03
||Name (click for home page)
||Jan 2004 - Dec 2006
||Research Staff Member, Storage and Data Services Research group, IBM T. J. Watson Research Center (Hawthorne, NY)
||Jan 2005 - Dec 2005
||Computer Scientist, Acrobat team, Adobe Systems (India)
||Dec 2005 - Dec 2007
||Software Engineer, Google Inc. (Mountain View, CA)
||Sep 2004 - Dec 2005
||Staff Engineer, Storage and Avaliability Group, VMware Inc. (Palo Alto, CA)
||Adam David Alan Martin
||Jan 2004 - Dec 2007
||Senior Software Engineer, Plaforms Team at MongoDB (New York, NY)