Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.
To get the free app, enter your mobile phone number.
The Practical Performance Analyst Paperback – October 31, 2000
The Amazon Book Review
Author interviews, book reviews, editors picks, and more. Read it now
Frequently bought together
Customers who viewed this item also viewed
Top customer reviews
There was a problem filtering reviews right now. Please try again later.
The core approach is Performance By Design, which is aligned to product development. His approach, if used properly, will ensure that performance goals are established in the design phase, and are met as a system or software evolves through the development life cycle.
Highlights of the book are:
(1) Through introduction to the foundation of performance: queuing, parallelism and multiprocessor systems.
(2) Coverage of contemporary issues, such as client/server and web system performance,
(3) Unexpected forays into performance characteristics and considerations that I've encountered in no other book. For example, Part 3 of this book addresses subtle issues such as transient analysis, scaling behavior and similar topics. Here the author integrates theoretical physics into performance analysis - while this may seem odd, it only reinforces that much can be added to the performance analysis body of knowledge by drawing from sources outside of computer science. His qualifications for this material includes a Ph.D in theoretical physics, and his ability to clearly explain concepts that are foreign to the average computer scientist or performance practitioner is excellent.
I like the conversation style that the author employs, the way he starts with the basics and builds upon them and the thoroughness in which all aspects of performance are discussed. More importantly, although advanced math concepts are introduced the way they are presented can be understood by anyone with high school or college freshman knowledge of probability and calculus.
The "universal scalability law" that he discusses extends Amdahl's Law via the addition of a "queuing" term that models effects such as data exchange between parallel processes. And although Dr. Gunther suggests that this queuing term ought to grow linearly with the number of parallel processes, this queuing term depends on the specific communication architecture of the computer system and can grow non-linearly, for example, as log to the base two of the number of processes.
This logarithmic growth law can occur because one processor may not communicate directly with all other processors. Instead, one processor may send information to two other processors, and each of those two processors may send information to two more processors, and so forth. Therefore, in order to model the communication that occurs in such a communication cascade, the queuing term should grow as log(n).
Moreover, performance data that are obtained from current parallel computer systems do not always conform to Dr. Gunther's "universal" scalability "law" under other conditions. For example, a large volume of data that exceeds the capacity of the total cache memory when distributed across a few processors may well fit into total cache memory when distributed across a larger number of processors. Under these conditions, the scalability for the larger number of processors appears to grow "super-linearly" relative to the scalability of a few processors. Thus, although Dr. Gunther's book is a useful introduction to the subject of measuring and modeling the behavior of parallel computer architectures, his universal scalability law should not be considered to be universal.
I found it interesting that he gives an example of typical computer time periods scaled up to human proportions. If a computer clock speed in nanoseconds were analagous to seconds then a main memory access of 100 cpu cycles would be like some minutes and a disk access would be like some months.
His coverage of queuing concepts is very accessible with a minimum of math.
Scalability is frequently discussed concept that often is not very well quantified. He has the most original approach to quantifying scalability that I have seen.
A queuing modeling package called PDQ is also provided with the book. The source code in C is provided for the PDQ package.
There are some PDQ examples within the book. This is a real bargain because certainly the PDQ software is worth much more than the cost of the book.
If you have and interest in capacity planning and performance analysis (especially if you work in this area) this is a must have book.
His writing style is easy to understand - not only does he give you the theoretical background, he shows the practical application and the results. He also includes code examples for the "PDQ" analyzer which is a software program he wrote which leverages his "Universal Law of Scalability." For more information, check out his web site at [...]
I highly recommend this book if you are a capacity or performance analyst and are interested in real-world application rather than just another boring queuing theory textbook.