Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: maxthr system statistic

Re: maxthr system statistic

From: Mladen Gogala <mgogala_at_verizon.net>
Date: Thu, 18 Jan 2007 23:20:27 -0500
Message-id: <1169180427l.4715l.2l@medo.noip.com>


On 01/18/2007 06:42:39 PM, Allen, Brandon wrote:
> Does anybody have a good understanding of how the maxthr metric is
> computed or how it's used by the CBO?

Interesting question. My assumption, which I was never able to prove, is that this, as well as the CPU speed is used to determine the relative cost of I/O and CPU cycles. As the holly book of Jonathan tells us, CBO converts everything to I/O blocks to read, CPU cycles included. It is my belief that throughput and CPU speed are used to calculate the relative prices. I don't know the exact algorithm, that you will have to ask Jonathan himself. He and Christian Antognini are the only people outside Oracle Corp. that fully grokk CBO.
Of course, praise also goes to Oracle Corp. for producing such rapidly changing, insufficiently documented and complex peace of sh...oftware that you need to be a genius to fully grokk. Kudos also go to Oracle for providing Oracle 11g only to the chosen few, who will profit by writing books and articles. The rest of us will thirstily try to get any drop of precious knowledge about the dearly paid software that Oracle would let us have. I may be forgetting my geography, but Redwood Shores, CA is beginning to look startlingly like Redmond, WA. Resemblance is too great for comfort.

-- 
Mladen Gogala
http://www.mladen-gogala.com

--
http://www.freelists.org/webpage/oracle-l
Received on Thu Jan 18 2007 - 22:20:27 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US