RE: Raid 50
- From: "MacGregor, Ian A." <ian@xxxxxxxxxxxxxxxxx>
- To: "'oracle-l@xxxxxxxxxxxxx'" <oracle-l@xxxxxxxxxxxxx>
- Date: Wed, 14 Jul 2004 10:31:54 -0700
While, I have enjoyed the frivolity, there is something to cosider, If =
Oracle cannot perform well enough on Raid5 systems, then it may be =
unsuitable for very large databases. When one has to store petabytes =
worth of data in both online and near-line storage, with hundreds of =
terabytes online, power consumption is a significant cost. Systems =
with the need to store exabytes worth of data are not far off.
This is not to say there are no performance penalties associated with =
using Raid 5 and Oracle. If CERN's Large Hadron Collider projects =
picks Oracle to store its event data, I'll bet they use Raid 5.=20
Ian MacGregor
Stanford Linear Accelerator Center
ian@xxxxxxxxxxxxxxxxx =20
-----Original Message-----
From: Mogens N=F8rgaard [mailto:mln@xxxxxxxxxxxx]=20
Sent: Tuesday, July 13, 2004 11:44 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: Raid 50
Why indeed stop with RAID-5 when you can do RAID-6 (aptly named for the =
factor six it incurs on small writes)?
Then, finally, we could have the RAID-666.
As Cary pointed out to be, the Law Of Bigger Numbers (LOBN) applies =
here
- both with respect to number of IO's and number of dollars spent.
James Morle had his birthday party this previous weekend, and while =
having one beer we decided to announce (jointly with the help of the =
BAARF Party members, perhaps?) the RAID-42 system soon, complete with =
an official-looking press release and all.
It would probably automatically produce competitive marketing papers =
from the big vendors, explaining that the simplicity of RAID-510 is to =
be preferred to this new, un-tested RAID-42 technology by a startup =
company called BAARF Unlimited.
We all know RAID-4 (some of you might be using it without knowing it, =
even!), but RAID-2 is less known - it's the first RAID-level that =
introduces parity disks. And I MEAN, literally, parity DISKS. It's =
beautiful. It must be a historic oversight that it's never really been =
used.
Mogens
Thomas Day wrote:
> <sarcasm>
> My mind immediately leapt to the RAID-555 technology; but why stop=20
> ther=3D e?
> The more you RAID-5 it the better the performance, right?
> </sarcasm>
>=20
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put =
'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
