>>>
>>>Yes, because Oracle relied on the existence of such in the
>>>real OS environments it ran on. As opposed to the sad joke
>>>Windoze was at the time. This however has changed. If
>>>anything, the "multi-threaded" servers of those databases -
>>>which never were such, just simple time-sharing round-robbin
>>>threads -
Oh boy. Something tells me you haven't had your hands in the
Informix DSA code as much as I have. The threads scheduler in
that database server was anything BUT "simple time-sharing
round-robin threads". It was a brilliant database-smart scheduler
and as long as it ran on a good OS, you very seldon saw more than
5% of your processor cycles lost to kernel mode...and that
even at extremely high IO rates.
--
http://www.freelists.org/webpage/oracle-l
Received on Tue Apr 18 2006 - 11:47:33 CDT
