Hot-Data Tracking Still Baking For The Linux Kernel

[tomato]

ZFS already does this through the ARC algorithm.

And ZFS lacks support for device removal from pool, defragmentation (in a FS that fragments files by design) and raid level migration...

Both file systems are far from "complete".

[johnc]

ZFS already does this through the ARC algorithm.

Which means Linux should have it by the year 2050.

Shortly after it's integrated into Windows 27.

[ryao]

ZFS already does this through the ARC algorithm.

[phoronix]

Phoronix: Hot-Data Tracking Still Baking For The Linux Kernel

A few months ago I wrote about hot-data tracking for the Linux kernel, a VFS feature that could be used by Btrfs and other Linux file-systems for delivering improved performance. Unfortunately the patch-set didn't make the new Linux 3.8 development cycle, but hot-data tracking is still being worked on for merging into a future Linux kernel release...

http://www.phoronix.com/vr.php?view=MTI1OTc