Re: SQL Server for Oracle DBAs

From: joel garry <joel-garry_at_home.com>
Date: Thu, 29 May 2008 17:41:14 -0700 (PDT)
Message-ID: <191841da-119f-46bc-b2c7-5a6eb272b715@x1g2000prh.googlegroups.com>


On May 29, 1:25 pm, "Tony Rogerson" <tonyroger..._at_torver.net> wrote:
>
> So Joel, as one of the people in the ORacle forums I actually respect do you
> not think that apply oracle ideaology to SQL Server is a good idea? Or, is

SP2-0734: unknown command beginning "not think ..." - rest of line ignored.

> it not better to do like I stated and don't try and apply oracle theory -
> instead make sure you understand how the SQL Server engine works which is
> where David's link comes in.
>

I think there are layers of understanding that all need to be applied. Then, there needs to be actual measurements taken of what is really going on, and instrumentation in the code to see it.

The original idea was that proper relational design would allow the database engine to optimize access, but the reality is shortcuts by all db vendors and most app designers make it necessary to understand the structure of each db and how it works.

Configuration is another layer below that.

In my opinion all the vendors could do more to document how it works. Hey, 5 minutes looking at downlinks from David's link I see the stuff about torn pages. It seems to be saying all SS buffers are based on 8K cache. Is that true? Boy, I'd really like to use other cache sizes like Oracle can. etc.

jg

--
@home.com is bogus.
Don't forget to mount a scratch monkey.
http://www.reuters.com/article/technologyNews/idUSN2832159720080529?pageNumber=3&virtualBrandChannel=0&sp=true
(but don't tell Jane Goodall)
word: ables
Received on Thu May 29 2008 - 19:41:14 CDT

Original text of this message