From a previous posting:

[Randy Grein]

Matt,
I've posted on this a number of times. I believe my article in  
NetworkVAR magazine (Where RAID Excels, Feb '97) is one of the first  
published instances where the issue is discussed. Since then I have  
formulated the problem in mathematical terms and informally published  
the results a number of times, here and elsewhere. The magazine is  
long out of print but if you're interested I have an electronic copy  
somewhere; contact me offlist. (I can't find the original published on  
the web anymore with google, but I did find several references to it  
in patent documents. Sad.)

FWIW the statement quoted should be that 3 drive RAID 5 WRITES are  
slower than a single drive, RAID 0, 1,  1,0, 5 and 6 have similar  
performance in multi-access server reads which confuses the issue.  
Caching controllers mask the problem somewhat but can create blockage  
problems when disk writes pile up.

Randy Grein, Master CNE, CCNA

On May 13, 2008, at 11:53 PM, Morgan Govender wrote:

>> From a previous posting:
>
> Raid 5 with 3 disks is slower than a single disk on it's own.  5 disks
> in a raid set is between 2 & 3 times as fast (to read) as a 3 disk  
> RAID
> 5
>
>
> Any references, tech papers to backup this? Not to convince me but
> rather my management?
>
> Please consider the environment before printing this e-mail.
>
> By receiving, reading or acting upon this email you will  
> automatically be bound by the terms of the  Email Legal Notice which  
> can be viewed at http://www.mf.co.za/Disclaimer/Pages/E-mailDisclaimer.aspx 
>   , or send a blank email to disclaim at mf.co.za to obtain a copy.
> Mutual & Federal Insurance Company Limited is an Authorised  
> Financial Services Provider
> -----
>
> _______________________________________________
> Novell mailing list
> Novell at netlab1.oucs.ox.ac.uk
> http://netlab1.usu.edu/mailman/listinfo/novell