Btrfs devel - Oct 2010 - Parity RAIDs in btrfs / RAID5 or RAID6

Hi all, 

I had a technical question about implementation of parity raid levels in btrfs,
like RAID5 and or RAID6.

Considering WAFL, the NetApp filesystem, each block is 4Kb. 
When we read 1 block, we have one 4Kb I/O on only one disk of the raid group.
So the raid group has a "random read I/O power" of all the disks it is
made from.
Of course, WAFL was made form the beginning in order to use a NVRAM, which helps
a lot.

ZFS, with raidz1/2/3, uses a different idea. One block can have different size,
and, in order to prevent the "write hole" issue, it was decided to
spread the block in all disks of the raid group.
Pros : we never have raid inconsistencies, even without NVRAM. 
Cons: in random read I/O, raidz1/2/3 are very pool performer. One raid group has
the "I/O power" of one disk.
Today, for a production storage solution, this can be a real issue. 

My question was about the direction btrfs had/will choose.
Of course, a solution like WAFL would be the best choice, in term of
performance.

If someone knows something about this ... 

Regards, 

Remi Urbillac
*********************************
This message and any attachments (the "message") are confidential and
intended solely for the addressees.
Any unauthorised use or dissemination is prohibited.
Messages are susceptible to alteration. 
France Telecom Group shall not be liable for the message if altered, changed or
falsified.
If you are not the intended addressee of this message, please cancel it
immediately and inform the sender.
********************************

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html