The Art of Computer Systems Performance Analysis: Techniques for Experimental Design, Measurement, Simulation, and Modeling [Hardcover]

R. K. Jain (Author)

Book Description

Publication Date: April 1991 | ISBN-10: 0471503363 | ISBN-13: 978-0471503361 | Edition: 1

The Art of Computer Systems Performance Analysis "At last, a welcome and needed text for computer professionals who require practical, ready-to-apply techniques for performance analysis. Highly recommended!" -Dr. Leonard Kleinrock University of California, Los Angeles "An entirely refreshing text which has just the right mixture of theory and real world practice. The book is ideal for both classroom instruction and self-study." -Dr. Raymond L. Pickholtz President, IEEE Communications Society "An extraordinarily comprehensive treatment of both theoretical and practical issues." -Dr. Jeffrey P. Buzen Internationally recognized performance analysis expert ". it is the most thorough book available to date" -Dr. Erol Gelenbe Universit? Ren? Descartes, Paris ". an extraordinary book.. A worthy addition to the bookshelf of any practicing computer or communications engineer" -Dr. Vinton G. Cer??? Chairman, ACM SIGCOMM "This is an unusual object, a textbook that one wants to sit down and peruse. The prose is clear and fluent, but more important, it is witty." -Allison Mankin The Mitre Washington Networking Center Newsletter

Editorial Reviews

Review

"At last, a welcome and needed text for computer professionals who require practical, ready-to-apply techniques for performance analysis. Highly recommended!" -Dr. Leonard Kleinrock University of California, Los Angeles "An entirely refreshing text which has just the right mixture of theory and real world practice. The book is ideal for both classroom instruction and self-study." -Dr. Raymond L. Pickholtz President, IEEE Communications Society "An extraordinarily comprehensive treatment of both theoretical and practical issues." -Dr. Jeffrey P. Buzen Internationally recognized performance analysis expert ". it is the most thorough book available to date" -Dr. Erol Gelenbe Universite Rene Descartes, Paris ". an extraordinary book. A worthy addition to the bookshelf of any practicing computer or communications engineer" -Dr. Vinton G. Cerf Chairman, ACM SIGCOMM "This is an unusual object, a textbook that one wants to sit down and peruse. The prose is clear and fluent, but more important, it is witty." -Allison Mankin The Mitre Washington Networking Center Newsletter

From the Inside Flap

As a key step in the selection of processors, operating systems, database, systems networks, and languages, performance analysis plays an integral role in the design and procurement of new computer systems. Presenting a unique blend of measurement, simulation, and modeling methods in a straightforward, problem-oriented fashion, The Art of Computer Systems Performance Analysis is a computer professional's treasure trove for practical performance analysis techniques. Integrating essential queueing theory with simulation, measurement techniques, data analysis, and experimental design, the book is divided into the following comprehensive parts:
* Overview of Performance Evaluation-addresses common mistakes in the evaluation process, criteria for selecting evaluation techniques and performance metrics, and the establishment of performance requirements
* Measurement Techniques and Tools-examines workload selection and characterization, monitors, capacity planning and benchmarking, data presentation, and ratio games
* Probability and Statistics-discusses the summarization of measured data, the comparison of systems using sample data, and regression models
* Experimental Design and Analysis-provides the maximum amount of information with the minimum number of experiments
* Simulation-clarifies the analysis of results, model verification, validation, and random-number generation
* Queueing Models-models and techniques for the solution of a wide variety of common performance problems
Analysis techniques are supported by more than 150 examples and case studies-all from the computer science field-and many are summarized with results in handy reference boxes. In addition, common mistakes in the application of various techniques are descussed, and a host of special tricks to demonstrate system superiority are revealed. The Art of Computer Systems Performance Analysis is an essential tool for performance analysts, designers, purchasers, and users of computer systems.

See all Editorial Reviews

Product Details

• Hardcover: 685 pages
• Publisher: Wiley; 1 edition (April 1991)
• Language: English
• ISBN-10: 0471503363
• ISBN-13: 978-0471503361
• Product Dimensions: 6.6 x 1.7 x 9.6 inches
• Shipping Weight: 2.5 pounds (View shipping rates and policies)
• Average Customer Review: 4.6 out of 5 stars  See all reviews (19 customer reviews)
• Amazon Best Sellers Rank: #247,043 in Books (See Top 100 in Books)
  • #9 in Books > Computers & Technology > Programming > Software Design, Testing & Engineering > Performance Optimization

More About the Author

Discover books, learn about writers, read author blogs, and more.

Most Helpful Customer Reviews

16 of 17 people found the following review helpful

5.0 out of 5 stars Lays the foundation April 21, 2002

By Mike Tarrani HALL OF FAMETOP 500 REVIEWER

Format:Hardcover

This seminal book was the most complete mainstream text on the subject when it was first published in 1991. The reason it remains popular eleven years later (which is an extraordinary life for a computer book) is because it's an applied mathematics book that focuses solely on performance analysis, simulation and measurements. Technology may change at a rapid pace, but the techniques for measuring it do not.

The book is broken into six parts consisting of 36 chapters and ten appendices. Part I is an overview of performance evaluation. I liked the chapter on common mistakes and how to avoid them, and the guidelines for selecting techniques and metrics.

Measurement techniques and tools are the focus of Part II. the eight chapters in this part address the full spectrum, including workload types, characterization techniques, capacity planning and benchmarking, and data presentation. Most of the chapters are generic enough to be timeless. The chapters covering execution monitors and accounting logs are showing their age, but the concepts are still valid. I especially liked the chapters about data presentation and ratios.

Part III is a refresher in statistics and probability, and can safely be skipped if your knowledge and skills are fresh. Part IV delves into experimental design and analysis, while Part V covers simulations. The remaining six chapters of the book, Part VI, are devoted to queuing models.

Note that although the math is clearly explained you need college level skills in order to fully comprehend the techniques presented. I recommend investing in MathCAD or a similar program to make it easy to work the exercises that end each chapter.

This book is one of the handful that will be on your bookshelf for years to come. It's probably one of the most frequently referenced works in the performance analysis, resource management and capacity planning knowledge domains. Because it has well designed exercises at the end of each chapter it's also well suited as an advanced college level text. In my opinion it's essential reading for performance and capacity analysts, and provides an excellent foundation for more specialized books, such as those by Menasce and Almeida, that address topics such as e-commerce performance, web services, etc.

11 of 12 people found the following review helpful

4.0 out of 5 stars Excellent - just buy it November 27, 2001

By A Customer

Format:Hardcover|Amazon Verified Purchase

I found this book by accident looking for some statistics help for a benchmarking project. It turned out to be worth its weight in gold as it is very clear, comphrehensive, and readable.

He outlines and elaborates a methodical, straightforward approach to performance analysis and provides excellent sections on what to do, not do, and how to validate or refute the analyses of others.

Extensive examples with solutions and enough equations to let you easily implement the analysis portions in code (or a spreadsheet if need be).

If you are doing any kind of performance analysis, whether it is for computer systems or not, you should get a copy.

The only reason I didnt give it 5 stars is a very small number of the example answers have typos or rounding errors. Be sure to get the errata list from the website.

10 of 11 people found the following review helpful

5.0 out of 5 stars a classic text.. May 18, 2002

By R. Arni

Format:Hardcover

This is an absolute gem. I really think that this book must be read by every CS/EE engineer. Performance analysis is always something that comes as an afterthought and engineers think that some some "mathematical" guy called a perf analyst will stop by after the product is completed and analyze the product. This is def wrong. PA should be done at every stage of production.

This book has been written assuming a novice reader. Several parts of the book have to re-read to really understand what the author is trying to convey, but trust me you will really appreciate it.

I suggest reading the following parts of the book(in order):

Part I (whole)
Part II (4, 5, 6, browse {7,8}, 9, 10 ) 10 is cool stuff..impress your peers with this
Part III (whole) read, re-read, re-read till you have digested every line..worth it really!!!!
Part IV (whole) read, re-read, re-read till you have digested every line..worth it really!!!!
Part V : You can skip this if you are not into simulation.
Part VI : Not really that easy to follow. I suggest Gunter's book for this.