The idea of a universal database (UDB) was invented as a reaction to the object-oriented databases of the early 1990s. Capabilities were added to the prevailing dominant design of relational database management systems (RDBMSs), including user-defined functions and data types, enhanced stored procedures and triggers (active database components), and SQL extensions.

Chamberlin’s comprehensive guide to the DB2 UDB is suitable for entry-level database developers as well as advanced technicians. It is a pleasure to find such a formidable subject treated in a readable and entertaining yet professional manner, with no assumption that the audience is intellectually challenged (as, for example, occasionally happens in the Dummies series).

The opening chapter contains a fascinating brief history of Structured Query Language (SQL), in which the author and his colleague Ray Boyce are deservedly credited with the invention of what has proven to be a powerful and compelling nonprocedural English-like interface to relational databases. SQL basics are covered in the next two-and-a-half chapters. Chapter 5, “Query Power,” looks at interesting SQL extensions, including new built-in functions useful in data warehousing (and decision support) applications, rollup, cube, and grouping sets. This illustrates a familiar feature of technology innovation and evolution: built-in functions that once required purchase of separate proprietary products such as Red Brick are now bundled as part of the UDB product. The tools for development of additional customizations are treated in chapter 6, on datatypes and functions; chapter 8, on Dynamic SQL; and chapter 9, on stored procedures (SPs). Be prepared to spend extra time with the details of these chapters, since they contain the technical details all developers crave and require. An especially engaging fact is that the approach to stored procedures enabled by IBM makes use of the procedural languages C, C++, a version of Basic, and (especially of note) Java as the host code in which the SP’s SQL is embedded. This sets a standard for openness, interoperability, and portability, which may make the other database vendors with proprietary scripting languages cringe. Convergence on Java is a possible outcome.

Finally, chapter 10 contains important details on database administration, backup and recovery, and another important technological innovation, interpartition parallelism. This brand of parallelism is to be distinguished from intrapartition parallelism on a symmetric multiprocessor (SMP) machine and is characterized as a shared-nothing or massively parallel system. Unlike SMP, the processors do not share memory, disk controller, or operating system software. If data from another node are required, a message is sent and a function is shipped. The nodes are connected by what amounts to a high-speed switch, including versions of Ethernet, ATM, or a proprietary IBM switching solution. Each processor executes its own image of the operating system--usually IBM’s version of UNIX (AIX)--on top of scalable parallel (SP2) hardware. As the author accurately explains, the data manipulation language (SQL/DML) is completely indifferent to whether the version of DB2 is what used to be called the parallel edition or the vanilla RISC System/6000. However, additional data definition language (DDL) syntax and redistributing operations are required to map the database partitions to hardware nodes and node groups. In fact, matters are a bit more complicated than the ten concise and well-written pages here suggest. But that is why there are several volumes of system documentation listed in Appendix F. Some assembly is required--and an additional software license. One can only wonder if the parallelism described here is the mainframe of the future. Even if it is an overly bold assertion to suggest that it is, especially given the dynamics of availability, Chamberlin’s engaging narrative lends powerful support to the claim that DB2 UDB is truly universal.

It is hard to imagine that in a superbly written book of nearly 800 pages, with six useful appendices, an excellent, detailed index, and supplementary exercises on the Web (www.mkp.com), anything more could be said. Even the ongoing Webification of the planet falls within the book’s scope, since dynamic SQL with Java receives a useful treatment with a sample program (pp. 498ff). The nicely edited and prepared text is supplemented with charming illustrations by Duane Bibby and a foreword by Don Haderle. For database professionals and developers accessing DB2 data through UDB, this text is likely to become a well-worn and dog-eared resource, making life in the software trenches a bit less difficult and a bit more like operating within view of an object-relational future, toward which so many projects are converging.