|
Posted by DA Morgan on 04/13/06 06:14
Robert Klemme wrote:
> DA Morgan wrote:
>> Robert Klemme wrote:
>>> Hugo Kornelis wrote:
>>>> - Oracle is more efficient. You have to spend less on hardware to get
>>>> the same performance.
>>>
>>> I'd put a question mark here - at least for the general way you put
>>> it. It always depends on what you do with the tool. SQL Server's
>>> network protocol seems to be more efficient as it allows faster round
>>> trip. This can make a difference in applications that touch the DB
>>> frequently accessing only small data sets.
>>>
>>> Kind regards
>>>
>>> robert
>>
>> Actually I've tested this myth in my lab and, so far have never been
>> able to duplicate it, given a decently tuned system. Do you have any
>> benchmarks that support it?
>
> I don't have figures at hand but it should be fairly easy to set up a
> test. All you need is JDBC drivers for both (I recommend jTDS for SQL
> Server), an installation of each product and then test timings for SQL
> that does nothing ("SELECT 1 FROM DUAL" for Oracle or "SELECT 1" for SQL
> Server) - you can even use a PreparedStatement to save the parsing. You
> don't even need to specially tune the db.
>
> Kind regards
>
> robert
That's what I've done. I've never seen it. But running SELECT 1 may be
more of a test of how clueless the person doing the test is than of the
hardware and operating system. As the two events are not at all similar
in what they do.
--
Daniel A. Morgan
http://www.psoug.org
damorgan@x.washington.edu
(replace x with u to respond)
Navigation:
[Reply to this message]
|