<< 1 >>
Rating: 
Summary: Up-to-date ... sends an old friend into semi-retirement
Review: Before this book was published my primary reference and personal favorite TP book was "Transaction Processing: Concepts and Techniques" by Jim Gray and Andreas Reuter. At over 1100 pages that book thoroughly covered the basics and drilled down into the nuances of transaction processing in a way unmatched by other books on the topic since it was first published in 1993. This book changes that by going far beyond transaction processing. It starts with the same fundamentals as the older book, and even covers many of the same topics, such as concurrency control, but it addresses each topic from a much wider perspective. For example, the discussion of concurrency goes far beyond the issues of transaction processing as a middleware component. It extends into application, database and search issues. Another indication that this book is more up-to-date is the material on queue managers. While they are at the opposite end of the spectrum from transaction processing monitors, they are integral to any discussion of transactional information systems. More importantly, both transaction processing monitors and queue managers are used in modern enterprise architectures. Having both topics discussed in great detail is a major point in this book's favor. Personally I intend to keep my copy of the older "Transaction Processing: Concepts and Techniques" because it does cover some of the subject matter more deeply. However, this book has replaced it as my principal reference and if I had to choose between them this is the one I'd go with.
Rating:
Summary: Up-to-date ... sends an old friend into semi-retirement
Review: Before this book was published my primary reference and personal favorite TP book was "Transaction Processing: Concepts and Techniques" by Jim Gray and Andreas Reuter. At over 1100 pages that book thoroughly covered the basics and drilled down into the nuances of transaction processing in a way unmatched by other books on the topic since it was first published in 1993. This book changes that by going far beyond transaction processing. It starts with the same fundamentals as the older book, and even covers many of the same topics, such as concurrency control, but it addresses each topic from a much wider perspective. For example, the discussion of concurrency goes far beyond the issues of transaction processing as a middleware component. It extends into application, database and search issues. Another indication that this book is more up-to-date is the material on queue managers. While they are at the opposite end of the spectrum from transaction processing monitors, they are integral to any discussion of transactional information systems. More importantly, both transaction processing monitors and queue managers are used in modern enterprise architectures. Having both topics discussed in great detail is a major point in this book's favor. Personally I intend to keep my copy of the older "Transaction Processing: Concepts and Techniques" because it does cover some of the subject matter more deeply. However, this book has replaced it as my principal reference and if I had to choose between them this is the one I'd go with.
Rating:
Summary: A wonderfully written book on an important topic
Review: Database concurrency control and recovery is one of pinnacles of computer science. An amazing collection of models, theoretical results, and implementation techniques enable thousands of users to simultaneously pound on a large database implemented on unreliable disks and networks, with full confidence that their data will be correctly stored. This book tells how this miracle is accomplished.
I teach database systems and also do research on databases, including systems-level refinements to concurrency control and recovery algorithms. This book has been invaluable to me in understanding the three major aspects of concurrency control in databases: the beautiful theory, the carefully constructed algorithms, and the specifics of the practice.
When this book first came out two years ago, I read most of it over a period of an intense week. That was such an enjoyable experience, because the book is very well structured and written in a smooth yet careful style. The authors ensured that all required concepts were in place before introducing a new concept. And the prose just flows, rendering difficult concepts understandable through well-chosen examples.
Since then I have referred to this book often with specific questions that arose in my research. Each time, my question has been answered fully in the book.
Each chapter ends with a section entitled "Lessons Learned" which summarizes the key ideas of the chapter and just as importantly, states the practical application of each concept. Some concepts have not yet been realized in practice; the authors are up front about this and explain why.
Mike Tarrani's review does a good job of explaining the similarities and differences between this book and the other seminal book on transaction processing, by Jim Gray and Andreas Reuter. Both books have their place, and both should be on the shelf (and read by!) all those who want to understand transaction processing at a deep level. And I agree with Jim Gray who noted in his foreword to the Weikum/Vossen book that it is likely to become (indeed, has) the standard reference in this field.
Rating:
Summary: Very Very Good
Review: Exceptionally clear writing. Encyclopedic in its coverage of transactions. Anyone dealing with transactions (not just DB developers) would find this a very valuable resource. If you need a book on transactions, this should be your first choice.
Rating:
Summary: A must for serious DB professionals
Review: Tradicionally, the 'transaction' concept is always discussed at the RDBMS classes in the University. But this book goes FAR BEYOND this wide known implementation; with a solid math foundation (some math required, specially set theory and algorithms ) it shows you there are a lot of 'real-life' transactions that requiere some formal methods for defining their implementation. This book has no discussion or topic regarding any comercial vendor technologies (specially databases), and I think this is very good. The Page and Object models for transaction processing are clearly explained. There's a very nice discussion concerning RAID technologies. This is not an 'academic' book in all the sense of the word. It can help IT professionals to make better transactional system desing (databases, workflow,e-business,etc). I would like some RDBMS vendors will include this kind of theory in their documentation....
<< 1 >>
| 
