Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle versus MS Sql Server
The usual mistake:
" Yes, in SQL and in application code I try to avoid using DB-specific
constructs. The DB-specific fine-tunig may come into place, if the database
is deployed on a server. Not changing SQL, of course. But using memory /
caching / ... techniques, which you probably know much better than I, in
the
database."
Databases differ from vendor to vendor. No they really do, they are not commodities. How they handle concurrency (some handle it well some poorly) , cursors (client side vs server side), and many many more. These are application architecture issues. I agree with Galen, you are designing an application to use the lowest common denominator and ignoring the dollars the customer spent on their database. My guess is you avoid using host variables. (which is going to make scalability go down the tubes and no DBA can fix.) I worked with someone who didn't think host variables made a difference. He was having a performance problem - 1,000 record processing benchmark was taking 26 minutes and pegging the server's CPU. We spent the couple of hours to convert to host variables and the same benchmark (same machine etc.) now took 6 minutes and averaged 10% CPU with 30% CPU usage peaks. (We also reused cursors and didn't close them so we only had to change the host variable values, rebind, reexecute.)
The whole system has to be done as a team. The people that wrote the database software usually put out a lot of application developer documentation (or they should) on how best to use their system, the application developer, the end user, the QA folks, the DBA, the Data Modeler (if your DBA doesn't do this) etc. are all needed to develop a high performance, scalable professional application.
Jim
"Michael G. Schneider" <mgs_at_mgs-software.de> wrote in message
news:9rdimi$l92$00$1_at_news.t-online.com...
>
>
>
![]() |
![]() |