Optimizing Oracle Performance

Chapter 7. Oracle Kernel Timings

Regardless of whether you access the Oracle kernel’s timing statistics through extended SQL trace data, V$ fixed views, or even by hacking directly into the Oracle shared memory segment beneath those V$ fixed views, the time statistics you’re accessing were obtained using a simple set of operating system function calls. Regardless of which interface you use to access them, those timing statistics are subject to the limitations inherent in the operating system timers that were used to produce them. This chapter explains those limitations and describes their true impact upon your work.

Operating System Process Management

From the perspective of your system’s host operating system, the Oracle kernel is just an application. There’s nothing mystical about how it works; it’s just a huge, extremely impressive C program. To gain a full appreciation for the operational timing data that the Oracle kernel reveals, you need to understand a little bit about the services an operating system provides to the Oracle kernel.

In this section, I am going to focus my descriptions on the behavior of operating systems derived from Unix. If your operating system is a Unix derivative like Linux, Sun Solaris, HP-UX, IBM AIX, or Tru64, then the explanations you will see here will closely resemble the behavior of your system. You should find the descriptions in this section relevant even if you are studying a Microsoft Windows system. If your operating system is not listed here, ...