If formally documenting your work is not your style, you can still follow this process on a piece of scratch paper. During my firefighting training courses, I don't have students create a document. Instead, on a piece of paper, they outline each key section and methodically perform their analysis. Within 30 minutes, the students can complete a full 3-circle analysis and be ready to discuss each circle's unique situation, how they relate, and possible solutions!

There are two distinct methods to diagnose Oracle contention. The traditional approach is based on performance ratios, such as the block buffer cache hit ratio. The current approach is based on Oracle's instrumented9 kernel code. The instrumented analysis approach is commonly called wait-event analysis. As you'll come to understand, wait-event analysis is far superior because it is both fast and accurate. Plus it is a key input when performing an Oracle response-time analysis. This book exploits Oracle's instrumentation. Here, I'll give you some insight into the wait-event analysis method. How to perform a thorough wait-event analysis is described Chapters 2 and 5.

What is the difference between traditional performance ratio analysis and wait-event analysis? I like to contrast the two by telling a story about a mother bringing her sick baby to a doctor. The doctor looks at Johnny and asks, "Johnny, where does it hurt?" Since Johnny is just a baby and can't talk, he looks at the doctor and starts screaming loudly. To diagnose little Johnny, the doctor must perform a series of uncomfortable tests.10 After the tests are performed, the doctor says something like, "Based on the series of tests I ran and based on my years of experience and my keen knowledge of a child's physiology, I believe this is the problem and this is what we should do." Very professional indeed!