Universal Meta Data Models
| 2004-04-09 00:00:00 | | 0 | Data Model
* The heart of the book provides the complete set of models that will support most of an organization's core business functions, including universal meta models for enterprise-wide systems, business meta data and data stewardship, portfolio management, business rules, and XML, messaging, and transactions
* Developers can directly adapt these models to their own businesses, saving countless hours of development time
* Building effective meta data repositories is complicated and time-consuming, and few IT departments have the necessary expertise to do it right-which is why this book is sure to find a ready audience
* Begins with a quick overview of the Meta Data Repository Environment and the business uses of meta data, then goes on to describe the technical architecture followed by the detailed models
User review
Rather superficial treatment of metadata management
If you want to know how to model certain types of metadata, this book might be of some help. It didn't help me to understand how to build a metadata management system. The book proposes a metadata management architecture; it describes the obvious parts of this architecture in great detail while the really interesting questions remain open. For instance, the really critical part is metadata integration. The book spends two meager paragraphs for this topic. The book repeatedly refers to metadata integration tools which can be bought or developed. A survey of such tools is however missing; thus if you want to know whether to buy or build a metadata management system, the book won't help you. Another example for missing detail is the statement that many architects treat the metadata management system as a data warehouse, while in Marco's view it is an operational system (which probably means that metadata should be continuously loaded), but the book never explains what an `operational` metadata management system would be (there are still metadata sources which are used in an operational way).
User review
Good reference from a RDBMS perspective
Although this is a well written book, for a book published in 2004 with the word `universal` in the title, I expected a little more modern treatment of metadata. This book begins with the statement that Mr. Marco is the `world's foremost authority on metadata`. And although I agree that the book does have several good points, it does not appear to be a well-rounded textbook on metadata. There are however several well-written sections on the relationship between the Capability Maturity Model (CMM) and metadata and the process of creating data stewardship teams. I was disappointed that the book does not even mention ISO/IEC 11179, RDF, OWL or any other modern tuple-based metadata technologies. The book does however mention how to use COBOL REDEFINES, DASD and other mainframe techniques to use metadata. The central thesis of the book is that metadata can and should be stored in a relational databases. The argument is that although relational structure are sometimes limiting, the prevalence of SQL skill sets make it practical to use relational databases to store metadata. The book includes a CD with ERWIN? files and a 30 day evaluation copy of ERWIN. The book also suffers from some formatting omissions such as the fact that most of the data models listed in the appendix lack titles.
User review
An Essential Resource for Managing Complexity
Universal Meta Data Models is filled with practical insights and world-class implementation guidelines gleaned from years of consulting on complex system integration and data warehouse projects. David Marco combines technical information with the bottom-line business case to roll it out.
Do you know where your data is? Meta data is structured knowledge management. It's about keeping track of what's happening in your IT systems -- databases, transformations, process flow, business rules, reporting -- and synchronizing that understanding across systems and teams. If you already know that and want to make it happen, this is the book for you.
If you are new to meta data, you will probably want to read Marco's prior work, Building and Managing the Meta Data Repository, first. Universal Meta Data Models picks up the meta data repository concept and extends it to the whole enterprise, in a Managed Meta Data Environment (MME). The MME has six components:
- Meta Data Sourcing Layer
- Meta Data Integration Layer
- Meta Data Management Layer
- Meta Data Repository
- Meta Data Marts
- Meta Data Delivery Layer
Marco describes each of these layers in detail, along with practical examples of how to roll them out in various industries. And just to spice it up a bit, he includes numerous morbidly humorous case studies illustrating the disasters that happen to companies that fail to properly managing their greatest asset: knowledge.
No comments:
Post a Comment