Tux3: the other next-generation filesystem

Posted Dec 3, 2008 8:38 UTC (Wed) by plougher (guest, #21620)
Parent article: Tux3: the other next-generation filesystem

Like any self-respecting contemporary filesystem, Tux3 is based on B-trees [...] Blocks are mapped using extents, of course - another obligatory feature for new filesystems

Of course this should be qualified as any self-respecting read/write filesystem. B-trees and extents are completely unnecessary for read-only filesystems.

Tux3 seems to have some nice design decisions which should offer high performance (reduced seeking). I like the variable sized inodes, (potential) optimised inodes for small files, and the packed attributes. Though I'm obviously bound to say that Squashfs has had variable sized inodes optimised for different file types/sizes for many years.