Analyzing Computer Systems Performance: With Perl: PDQ

Features

• Cover Type: Hard Cover with 436 pages
• Published by: Springer
• Edition: 1st Edition June 28, 2005
• Written in: English
• ISBN 10 Number: 3540208658
• ISBN 13 Number: 978-3540208655
• Book Dimensions: 9.4 x 6.4 x 1.2 inches
• Weighs: 1.8 pounds

Product Review


From the reviews:

"The literature of computer performance analysis is generally composed of two groups: heavy theoretical treatises and performance cookbooks. … Surprisingly, this book is exceptionally well balanced between theory and practice … . I strongly recommend this book, both for the novice practitioner and for the experienced performance analyst. Both can extract a vast array of benefits, ranging from understanding the theoretical concepts of performance modeling, to building for themselves a powerful modeling tool … ." (Jair Merlo, Computing Reviews, May, 2005)

Product Description


Analyzing computer system performance is often regarded by most system administrators, IT professionals and software engineers as a black art that is too time consuming to learn and apply. Finally, this book by acclaimed performance analyst Dr. Neil Gunther makes this subject understandable and applicable through programmatic examples. The means to this end is the open-source performance analyzer Pretty Damn Quick (PDQ) written in Perl and available for download from the author’s Website: www.perfdynamics.com.

As the epigraph in this book points out, Common sense is the pitfall of performance analysis. The performance analysis framework that replaces common sense is revealed in the first few chapters of Part I. The important queueing concepts embedded in PDQ are explained in a very simple style that does not require any knowledge of formal probability theory. Part II begins with a full specification of how to set up and use PDQ replete with examples written in Perl. Subsequent chapters present applications of PDQ to the performance analysis of multicomputer architectures, benchmark results, client/server scalability, and Web-based applications. The examples are not mere academic toys but are based on the author's experience analyzing the performance of large-scale systems over the past twenty years. By following his lead, you will quickly be able to set up your own Perl scripts for collecting data and exploring performance-by-design alternatives without inflating your manager’s schedule.

Reader Reviews
This book lays out basics of queuing theory in a way that is both understandable and easy to apply. I have fumbled for years using complex performance analysis tools without ever really knowing whether or not I'm getting the right answer. Now I know what looks right and what doesn't. Gunther has an easy to follow style as he introduces the concepts of queuing theory. Then he shows you how to apply what you have learned using real life examples. It's a definite "must have" for the capacity planner or system performance analyst.