Unioning file systems: Architecture, features, and design choices [LWN.net]

Unioning file systems: Architecture, features, and design choices

Posted Mar 20, 2009 17:26 UTC (Fri) by arnd (subscriber, #8866)
In reply to: Unioning file systems: Architecture, features, and design choices by masoncl
Parent article: Unioning file systems: Architecture, features, and design choices

To take this thought a little bit further, this could also incorporate:

* laying out files and metadata for performance characteristics of optical
media (seek times vary a lot with location, 2kb sector boundary, ...).

* compressing all files in advance may give you much better placement
options than online compression.

* This could be integrated into mkisofs in the same way that HFS support
is. Iso9660 leaves a significant amount of free space in the front, so it
may just fit.

* In a similar way, it could be done in the same way that ext3 conversion
works, but on an existing iso9660 file system to create a hybrid file
system. Bonus points for making it work with multisession writing on an
already burnt iso9660 disk.

(Log in to post comments)