Re: ASSM vs. non-ASSM

From: joel garry <joel-garry_at_home.com>
Date: Fri, 27 Jan 2012 09:36:59 -0800 (PST)
Message-ID: <4bc48304-0d53-4aac-a970-752a9660513b_at_nf9g2000pbc.googlegroups.com>



On Jan 27, 1:09 am, mhoys <matthias.h..._at_gmail.com> wrote:
> On Jan 27, 9:29 am, "Gerard H. Pille" <g..._at_skynet.be> wrote:
>
> > onedbguru wrote:
>
> > > Yes, but the next time you need "more space" for that index it is
> > > already allocated and will be used.
>
> > This is false.  The indexes keep growing for the same amount of data and queries using these
> > indexes become too slow.
>
> I would like to answer with "it depends" :-) Btw, I tried to develop a
> test case to show the performance difference between "normal" and
> "soufflé" B-tree indices, but I haven't been successful so far. This
> is on 10.2.0.5, ASSM tablespace with automatic extent allocation...
> What are the properties of your system, Gerard?
>
> Matthias Hoys

Also see http://jonathanlewis.wordpress.com/category/oracle/infrastructure/assm/ for some more ideas and background. Having a bunch of uncommitted inserts seems very interesting, it spotlights the problems with concurrency and tracking blocks with bitmaps.

jg

--
_at_home.com is bogus.
http://www.newsfactor.com/news/SAP-Sets-Database-Sights-on-Oracle/story.xhtml?story_id=1220020SU1PU&full_skip=1
Received on Fri Jan 27 2012 - 11:36:59 CST

Original text of this message