RE: normalization

  • From: "Steve Adams" <steve.adams@xxxxxxxxxxxx>
  • To: "'Ryan'" <ryan_gaffuri@xxxxxxxxxxx>
  • Date: Wed, 2 Nov 2005 15:25:51 +1100

Hi Ryan,

Just test to scale and measure it.
Of course, CPU usage is affected by many things, so it will not be easy
to express the results in a way that might be meaningful to management.
Nevertheless, in cases like this, I think that showing a non-trivial
difference is enough.

@   Regards,
@   Steve Adams
@   http://www.ixora.com.au/         - For DBAs
@   http://www.christianity.net.au/  - For all 

-----Original Message-----
From: Ryan [mailto:ryan_gaffuri@xxxxxxxxxxx] 
Sent: Wednesday, 2 November 2005 3:17 PM
To: steve.adams@xxxxxxxxxxxx
Cc: 'Oracle-L Freelists'
Subject: Re: normalization


Steve,

Do you know a way to take the data you presented and do a test to show how 
much additional CPU usage this will require under specific conditions?

if the trigger is fired 'X' number of times in a certain period, how much 
increased CPU will I need?
How much will this reduce response time and/or scalability?

If I was to present this to a manager, I would typically need to show 
something more tangible. Just because certain data elements that I gather 
from the data dictionary increase, does not mean it will cause a problem for 
us. So I would need to show realistic estimates.

any ideas on how to do this type of test?

--
//www.freelists.org/webpage/oracle-l


Other related posts: