Re: Oracle performance C OCI vs. JDBC thin?

Hi there,

Davide Bianchi wrote:
>
> "Pasi Kovanen" <Pasi.Kovanen_at_removethis.iki.fi.invalid> wrote in message
> news:a54sjo$8m0$1_at_news1.song.fi...
>
> > Has anyone compared (preferrably on Solaris
> > but Windows is OK) performance of queries using
> <zap>
>
> I did something like that, and my results are that the OCI
> driver is 33% faster than the Thin driver, but the OCI is
> also dependent on the Oracle Client installed on the machine,
> so you have another piece to install and configure.
>

An interesting comparison can be found at:

http://www.oreilly.com/catalog/jorajdbc/chapter/ch19.html

In this sample chapter Donald Bales (also referred to by 'AV') claims the Thin driver to be faster. Interestingly enough he describes only insert statements in his testing (though with reasons).

As a rule I try to test in a setup that is as close as possible to the actual application environment to be developed. This will often mean testing on larger and more complicated statements. I have recently done that for the application we develop and have found the OCI driver roughly 10 % faster than the Thin driver. This was done from a Windows 2000 client connecting to an Oracle 8.1.7 database running on HP-UV 11.0.

Another interesting point is that the Thin driver requires a TCP/IP connection, where the OCI driver works with other network protocols as well.

Regards,

Ruud de Koter

-- 
--------------------------------------------------------------------------------------
Ruud de Koter                    HP OpenView Software Business Unit
Senior Software Engineer         IT Service Management Operation
Telephone: +31 (20) 514 15 89    Van Diemenstraat 200  
Telefax  : +31 (20) 514 15 90    PO Box 831
Telnet   : 547 - 1589            1000 AV  Amsterdam, the Netherlands
Email    : ruud_dekoter_at_hp.com

internet: http://www.openview.hp.com/itsm
          http://www.openview.hp.com/assetview
intranet: http://ovweb.bbn.hp.com/itservicemanager
--------------------------------------------------------------------------------------