Re: A kind of... benchmark?

Holger Baer <holger.baer_at_science-computing.de> wrote

> Although I generally agree with your statement, I believe there is
> actually one thing where benchmarks are usefull: Learning the database.
<snipped>

Yes. What is great is to have a machine to muck about with, that is of the same approx size of production boxes, carrying the same volume. Then you can see compare using an I/O slave pool versus normal config. What happens in you use an obscene amount of memory for db block buffers. And so on.. :-)

But I would not really classify that as benchmarking. Learning/experimenting are much better terms.

> But then, this is not what the OP was after.

Yeah.. having a standard db metric (similar to whet/drystones or bogusmips) for a db environment is IMO almost impossible. Simply because Oracle performance is not just a function of the hardware, but that of good usage and proper design. Add to that the vast differences between OLTP vs OLAP vs batch vs ODS etc. etc..

--
Billy