Re: Designing database tables for performance?

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Mon, 19 Feb 2007 13:39:30 GMT
Message-ID: <mMhCh.8261$R71.127665@ursa-nb00s0.nbnet.nb.ca>

Walt wrote:
> "Bob Badour" <bbadour_at_pei.sympatico.ca> wrote in message
> news:KehCh.8254$R71.127461_at_ursa-nb00s0.nbnet.nb.ca...
>

>>Timasmith wrote:
>>
>>>On Feb 18, 8:00 pm, "Mike Preece" <mich..._at_preece.net> wrote:
>>>
>>>
>>>>On Feb 9, 5:04 am, "Timasmith" <timasm..._at_hotmail.com> wrote:
>>>>
>>>>Are you interested in logical or physical performance?
>>
>>Oh my! Can anyone imagine anything more embarassing than Mike's
>>question? Should we all point and laugh now?

>
> Two points.
>
> First, Mike's a Pickie. Enough said.
>
> Second, the term "logical vs physical" has been used in many different
> ways by many different respectable people in the industry.

Respectable? I doubt that.

In the
> documentation for one ancient DBMS, a so-called "logical I/O" was a
> reference to a buffer in memory,
> while a "physical I/O" was a retrieval of a buffer's worth from disk.

Regardless, both accesses are physical accesses: a physical access to memory or a physical access to disk followed by a physical access to memory. The two have different physical performance characteristics.

Performance is a consequence of physics not of logic. Received on Mon Feb 19 2007 - 07:39:30 CST