Oracle DBAs and developers are all too familiar with the outlay of time and resources, blown budgets, missed deadlines, and marginally effective performance fiddling that is commonplace with traditional methods of Oracle performance tuning. In Optimizing Oracle Performance, Cary Millsap, former VP of Oracle's System Performance Group, clearly and concisely explains how to use Oracle's response time statistics to diagnose and repair performance problems. Cary also shows how "queueing theory" can be applied to response time statistics to predict the impact of upgrades and other system changes. The price of this essential book will be paid back in hours saved the first time its methods are used.
Optimizing Oracle response time is, for the most part, a solved problem. I hope that I have written effectively enough that, after reading this book, this idea will come to fruition in your own experience.
However, if you're like most people, you probably don't feel that way yet. For most people, Oracle performance improvement projects are long, frustrating battles against some invisible enemy that evades detection no matter how much time or extra computer hardware you throw at the situation. The root cause of the problem is that most education about tuning is broken. My aims in this book are to show you why, and to reveal to you what you should do instead.
Bad tuning methods have prospered among the Oracle community for a long time. For well over a decade, the Oracle community has been afflicted with lots of performance problems but a virtual absence of competent training programs for performance analysts. The result has been a lucrative seller's market for the Oracle performance analyst. Throughout the 1990s in many parts of the world, a consultant could name his own price and bill by the hour for time spent attempting to improve performance. The tuning methods that evolved in this environment were geared more toward maximizing a consultant's revenue intake than maximizing the success of your system.