[ILUG] Why RAID

[Paul Jakma]

On Tue, 13 Jul 2004, Paul Mc Auley wrote:

> likely to recover from a serious disk failure, and doing stuff in 
> the main CPU rather than on a dedicated controller is going to 
> generate overheads.

CPU overhead is negligible. Indeed, it can be a win for RAID5, an 
Athlon or any other modern CPU will beat the pants off any of the 
CPUs used on mid->high end RAID controllers.

The overhead for host-based RAID is mostly bus overhead. You have to 
transfer data n times to controller(s) rather than once.

Apparently the Promise SX4 SATA controller is flexible enough to be 
able to solve this problem. From what I gather from l-k, it's sort of 
a RAID "accelerator" rather than an actual RAID controller - XOR 
engine and memory but leaves the host to drive RAID logic. So once 
Jeff Garzik and Neil Brown work out how to support this with 
linux-raid it will be the controller to have ;)

regards,
-- 
Paul Jakma	paul at clubi.ie	paul at jakma.org	Key ID: 64A2FF6A
 	warning: do not ever send email to spam at dishone.st
Fortune:
Q:	How many marketing people does it take to change a lightbulb?
A:	I'll have to get back to you on that.