Slow systems are frustrating. They waste time and money. But making consistently great decisions about performance can be easy, if you understand what's going on. This book explains in a clear and thoughtful voice why systems perform the way they do. It's for anybody who's curious about how computer programs and other processes use their time and about what you can do to improve them.

Through a mix of personal vignettes and technical use cases, Cary Millsap reviews the process of improving performance and provides best practices for optimizing systems efficiently. You'll learn how to identify the information needed to improve a system, how to find the root causes of performance issues, and how to fix them. You'll also learn how performance optimization is both a skill set and a mindset, and how to develop both over time.

If you're a computer professional whose success relies on software that goes fast, by the end of this book you'll be able to identify, view, scope, analyze, and remedy performance issues with consistency and confidence.

Table of Contents

Part I. Look at It

Part II. Method

Part Ill. Profiling

Part IV. Measuring Performance

Part V. Optimizing

Part VI. Delays

Part VII. Waste

Part VIII. Problem Solving

Part IX. Predicting

Part X. Latency Hiding

Part XI. Fallacies

Part XII. Testing

Part XIII. Planning

Part XIV. Politics

Part XV. Just for Fun

Making things faster is about making things better. Making life better. The faster your tools can work, the more time you have for doing the things you want to be doing. If your tools help you process information, then the faster they work, the better decisions you can make.

Optimizing—whether a computer, or a plow, or anything in between—requires two separate skill sets: one, asking the right questions; and two, of course, answering those questions. Most people who think about optimizing are well aware of this second skill set. Fewer seem aware of the first one. But this first one—the mere act of asking the right questions—is the skill you need to develop first, especially if you’re a leader. You might be surprised at how quickly you’ll be able to learn it.

There’s a problem, though. The questions I’ll teach you to ask are simple. But anyone who’s ever met an inquisitive child knows that simple questions aren’t always easy to answer. For example, the people who look after your company’s computer may know exactly what the system’s CPU utilization looks like at 2:00 p.m. on a given Friday, but they may not be able to answer how long it takes a clerk to enter an order. There’s a mismatch between the answers they’ll want to give and the questions you’ll want to ask. This mismatch itself hides opportunities to make things faster. I’ll teach you how to find them.

Optimizing is often more political than technical. It’s curious, then, that books about optimizing are almost always only technical. It’s easy for technical people to regard the nontechnical elements of a project as unnecessary and maybe even ridiculous disruptions. But the fact is, the nontechnical aspects of optimizing require understanding and effort, just like the technical aspects do. To optimize something, sometimes you have to confront the monsters created by panic and fear, in addition to being productive technically. A strong track record of successes will help you win debates about what to do next, but you’ll never be able to create that track record without exercising some political savvy.

My goal with this book is to help you improve both kinds of skills: the asking and the answering, the political and the technical. The payoff potential is huge. When you understand both the science of performance and the art of navigating the desires and emotions of the people who care about it, then you can optimize anything.

About the Author

Cary Millsap spent the 1990s learning a lifetime's worth of technical and political lessons as a consultant for Oracle Corporation. By 1999, he had amassed an impressive personal track record with about a hundred different customers, while building and training a team of 85 elite performance specialists that Oracle deployed to help hundreds more customers worldwide. In 2004, he wrote Optimizing Oracle Performance, for which he and researcher Jeff Holt were named Oracle Magazine's 2004 Authors of the Year. In 2008, he founded Method R Corporation, which served ultra-high-profile customers and created the world's first software system for managing, mining, and manipulating thousands of Oracle trace files. Over his career, Cary has educated thousands of information technology professionals through his commitment to writing, teaching, and speaking at public events.