The new methodology for performance and troubleshooting
                                                 Last update (2012-09-07 09:00:19)
                                                                                                                   Date added (2008-02-26 13:59:39)

As the years pass the DBA is in the pursuit of optimal database performance. The old method (which i never understood) is calculating system ratios about buffer hit, dictionary cache hit, library cache hit, etc.

But what can really mean for the application a 98% of cache hit, when the query takes ages to finish? Can you really convince someone about the system performance because you have a very high cache hit ratio while the user pressed the button and still waits?

The solution to the performance problems is simple. The first concept is that you can identify for a specific problem the exact reason for waiting and second that every problem is usually different from the others.

Finding where you are "waiting", means finding the root cause for poor performance.

Oracle records every wait at some tables, all this interface tables called Oracle Wait Interface(OWI). The wait interface of Oracle was introduced many versions ago, and with each new release, Oracle has added enhancements to the wait interface that helps DBAs better understand where database bottlenecks and waitings are occurring.

So learning to "translate" the Oracle Wait Interface as a DBA means you can troubleshoot every performance problem!

Other links
Current waiting events