|
5 of 6 people found the following review helpful:
5.0 out of 5 stars
Advice from real experts who are actually succeeding in DW, October 4, 2003
Too many professional books are written by "experts" with only a shallow understanding of what they're writing about."Secrets of the Best Data Warehouses..." is a happy exception. I'm a market analyst who tracks the telecom IT industry. I know one of the authors, Rolf Hanusa, has been the technical architect and driving force behind SBC Communications' enterprise data warehouse since 1993. And believe me, 10 years is an incredibly long time in the IT hot seat, especially considering the many failures in the telecom warehouse space. In fact, the SBC Communications data warehouse is today's largest telecom warehouse with 10s of thousands of users and 35 terabytes of on-line data. Not only is the warehouse handling an average 100,000 queries a day, but it's steadily taking over the reporting functions of operational systems such as billing, provisioning, and finance. It's one of those rare cases where the textbook "one-version-of-the-truth" warehouse is actually working. The "secrets" Hanusa et al. reveal are not technical tricks, but the guiding principles that make for a successful data warehouse implementation and on-going program. If you're a technical guy looking for speeds and feeds, system diagrams, and hardware specs, this book is not for you. Instead, it's written for the executives and technical managers whose jobs are on the line when a data warehouse fails. The book is written in plain English. Caricatures and analogies break up the text into bite-sized kernels. While it's an easy read, the book doesn't skirt the tough subjects. You'll get advice on things like: ensuring data integrity, handling the "query from hell", cost justifying the warehouse, central vs. decentralized structure, detailed vs. summarized data, and winning executive support. Several of the reviews I read suggest the book is biased towards the Teradata warehouse. This is hogwash. Teradata is not even mentioned in the book, but the book *does* take the concept of a centralized data warehouse to heart. In other words, don't buy this book if you're building a case for a stovepiped data mart. This book advocates just the opposite.
|