Re: Filesystem vs. Raw IO

From: Carel-Jan Engel <cjpengel.dbalert_at_xs4all.nl>
Date: Fri, 28 May 2004 23:01:05 +0200
Message-Id: <6.0.1.1.0.20040528225413.44d487b0@pop.xs4all.nl>

IBM claims they can reach nearly RAW-partition performance on filesystems using concurrent IO supported by JFS2. Some figures (from the paper http://unix.ittoolbox.com/browse.asp?c=UNIXPeerPublishing&r=http%3A%2F%2Fwww%2D1%2Eibm%2Ecom%2Fservers%2Faix%2Fwhitepapers%2Fdb%5Fperf%5Faix%2Epdf):

avg Trougput (tps) Raw LV's : Direct I/O : Concurrent I/O = 710.19 : 219.69 : 652.12.
avg Response Time (sec) Raw LV's : Direct I/O : Concurrent I/O = 0.06 : 0.12 : 0.07

They did the test with 9i. The paper also discusses the use of filesystems in database environments.

I hope to be able to perform some tests this year. I'm curious how this will compare with ASM in 10g.

Regards, Carel-Jan

===

If you think education is expensive, try ignorance. (Derek Bok)
===

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Fri May 28 2004 - 15:58:46 CDT