The key to client/server computing.

Transaction processing techniques are deeply ingrained in the fields of
databases and operating systems and are used to monitor, control and update
information in modern computer systems. This book will show you how large,
distributed, heterogeneous computer systems can be made to work reliably.
Using transactions as a unifying conceptual framework, the authors show how
to build high-performance distributed systems and high-availability
applications with finite budgets and risk.

The authors provide detailed explanations of why various problems occur as
well as practical, usable techniques for their solution. Throughout the book,
examples and techniques are drawn from the most successful commercial and
research systems. Extensive use of compilable C code fragments demonstrates
the many transaction processing algorithms presented in the book. The book
will be valuable to anyone interested in implementing distributed systems
or client/server architectures.

Average Customer Review: ( 9 customer reviews )
Write an online review and share your thoughts with other customers.

In particular, this book covers the following topics in more depth than the newer boom cited above:
- Fault tolerance and availability, both topics are covered in depth from hardware and software perspectives. This is unique for a book on transaction processing in that most books on the subject confine their scope to software and databases.

- A wide and complete survey of transaction models. True, some of this material is about models that are falling into disuse, but the value is the way the authors go deeply into the mechanics. I've always felt that this part of the book is the most valuable because the principles can be refactored into hybrid models. Moreover, comparing this material with the newer book by Weikum and Vossen shows that these principles are still employed in today's TP solutions.

Material about transaction processing monitors is obviously out of date, but, like the TP models, the principles still apply to contemporary systems. My recommendation is if you are going to buy a single book on the topic get the Weikum and Vossen I cited in the first paragraph. However, if your budget allows, I also highly recommend this book as well because of the depth in which fault tolerance and TP models are covered. If you want to just learn the basics of TP I recommend that you consider "Principles of Transaction Processing" by Philip A. Bernstein and Eric Newcomer because it is less daunting than this or the Weikum and Vossen book (both of which are 1100+ pages).

19 of 19 found the following review helpful:

The bible of transaction processingNov 22, 2001

I used this book as a CS grad student in college
for a class on transaction based systems and it covers
how to do transactions from top to bottom. Although
it was published in 1993 the techniques described in this
book are actually more advanced than techniques
used in a lot of real world systems today so it is not
out dated. I have yet to see a book as comprehensive as
this on how to actually implement transactions. Good
book for software engineers to read. My only complaint
is that the book has a lot of typos and some bugs in the
source code listings. Also because the book is so damn big
(i.e. lot of pages) they chose to use very thin paper which
makes it not very good for using hilight markers on. Still
this is the definitive book on how to implement transaction
processing.

15 of 15 found the following review helpful:

First and foremost, an implementor's bookJul 31, 1997

I know I wouldn't have fully assimilated even half of this book in school: it certainly provides overview material approachable by those of moderate experience, but it will be most appreciated by those who have actually implemented (or are about to implement) much of what it discusses.

Working on a hot new RAID box? Parts of this book will help. How about a file/record/object system? Yup. In either case, more complete references exist (and you'll need them), but the point is that this book really does cover data management from the disk on up in a useful degree of detail.

On the other hand, if you're implementing a transaction manager, this book is both necessary and likely sufficient. Read it first to get started in the right directions, and revisit specific areas as you code - a casual comment is easy to forget if you read it before you fully appreciate its context. The authors acknowledge that there are neat tricks not covered (because they would obscure the core material) that you'd likely want to incorporate once you're sure you really understand their impact, but all the framework, and much of the detail, is included.

Bottom line: If your interest in data management is truly intense, you need this book - it should remain current for quite some time.

15 of 16 found the following review helpful:

Very complete!Sep 10, 2000
By Dan Crevier
I wanted to learn more about transaction processing and had found bits and pieces from various sources. Well, once I got this book, I realized it was the only one I needed. It covers everything in a simple to read manner with lots of details (and good references to find more as needed). I highly recommend this if you want to learn more about transaction processing

17 of 19 found the following review helpful:

a must not only for the database implementation studyApr 03, 1999
By Ulrich Graser
If you read the book you have a clear understanding about the functionality and interactivity of the world's largest distributed database and transaction systems on different levels of abstraction. So the demands of the hardware architecture and the concepts of the software design from operationg system to application level are omnipresent illustrated and result in a feeling the reader will sense as a whole model of understanding.

I recommend the book to everyone who makes decisions about serious computer questions in databases and in general operation systems. It gives also a profound example for good software engineering. By the way it demonstrates how to write a reliable (!) OS, how to propagate the certainty of having a process or not from the clock dependent processor tic(k) level to the front end ATM human computer interface.

I hope there will be a next edition with updated values about the actual hardware specifications.

See all 9 customer reviews on Amazon.com