A field of research has reached a certain level of maturity when the first true textbook on the subject appears. This book is the first textbook on database machines. Ozkarahan, as a pioneer in database machine research, is well qualified to write this book and has done an excellent job. The book succeeds admirably both as a reference on database machine research and as a textbook for university courses.

Briefly, the chapter contents are described below.

• Chapter 1 introduces the rationale for database machines. Unique hardware and software architectures are needed to provide greater efficiencies for nonnumeric processing.

• Chapters 2 and 3 provide a comprehensive overview of database management knowledge.

• Chapter 4 presents the ideas of parallelism and pipelining. Implementations and the potential speedup associated with these techniques are discussed.

• Chapters 5 and 6 discuss associative memories and associative processors, respectively. Various implementations are described and compared.

• Chapter 7 is an extensive survey of database machine designs and implementations. These are divided into categories of cellular associative systems, multiprocessor-based systems, in-stream and/or pipeline processing, logic-enhanced primary memory (VLSI) systems, filters, and commercial database machines. The unique features and concepts of the systems are highlighted.

• Chapter 8 presents a classification scheme for database machines. A database machine system is described by the following eight attributes: (1) degree of parallelism (e.g., none, low, high), (2) associativity (e.g., none, partial, full), (3) processor organization (e.g., pipeline, multiprocessor, cellular), (4) type of processor coupling (e.g., SISD, MISD, SIMD, MIMD), (5) processor-memory coupling (e.g., static, dynamic/switched), (6) memory level (e.g., primary, secondary), (7) memory search (e.g., direct, indirect), and (8) host coupling (e.g., tight, loose, standalone). Using this classification scheme, a database machine theory is proposed.

• Chapter 9 discusses the software needed to provide database machine functions.

• Chapter 10 covers the role of database machines in distributed database systems.

• Chapter 11 handles the issue of database machine performance evaluation.

• Chapters 12 and 13 cover the basic issues of document and text retrieval. The application of database machines to unformatted, textual databases is used to demonstrate linkages between database systems and information retrieval systems.

Writing a textbook on database machines presents a special problem because of the newness of the field. While most textbooks provide a deductive progression of principles and theory into development and practice, it is still too soon for database machine theory to be fully discovered. Thus, the author has chosen a more inductive approach of instruction. Chapters 1 through 6 provide a solid background of fundamental concepts on database management and computer architecture. Chapters 7 and 8 form the center of the book. Using the individual cases of database machines surveyed in Chapter 7, a theory of database machines is proposed in Chapter 8. This theory is composed of the aforementioned classification scheme and a discussion of the key issues of database machine research. These issues include architecture optimization, data models, database filtering, database partitioning, exploiting locality, concurrency, and tagging. While these discussions are insightful, the author admits to only having scratched the surface of providing a complete database machine theory. Chapters 9 through 13 complete the book by presenting special topics of database machine research. Each chapter includes a set of exercises that are very well designed for student use.

I can highly recommend this book. It is an excellent reference to the database machine research literature. If I were to teach a course on database machines, this would be the text that I would use.