Do we cast the blame on the IO subsystem? While no one would be proud of a 22.5 ms IO subsystem response time, based on the 3-circle analysis, and unless there is a gross misconfiguration, performance optimizing solutions exist from Oracle, application, and operating system perspectives. Don't be quick to point the finger at the IO subsystem. And for you SQL tuners, don't immediately cast the blame on the SQL either, or you will be limiting your optimization opportunities.

Now that Oracle has instrumented its code, given the system calls convenient Oracle kernel developer names, stored the timing information, and made the information available in performance views, we are ready to query the wait time through the SQL interface.

Oracle makes its kernel instrumentation details available to us very simply through the SQL interface. There is actually a family of wait event views. I will present most of them in this book, starting with the three wait event views that form the core of any wait-based analysis and the queue time component of an ORTA: