1 Introduction Parallel Database and Knowledge - Base Systems

Database technology has been extended into many new applications areas such as computer-aided design and manufacture, geographical information systems, and real-time intelligent decision making. Such applications bring with them demands for richer information models and the capability to handle larger quantities of heterogeneous data. Furthermore , increasing numbers of online users are expecting some form of smartness from their information systems. We use the term smart information systems, or knowledge-base systems, to describe computer applications that incorporate nontrivial features such as the ability to adapt and evolve, or to carry out inferencing. All these factors combine to produce a need for improved performance and additional functionality. This book consists of 13 invited papers, chosen to illustrate ways in which novel parallel hardware is being used to provide improved performance and additional functionality for a variety of information systems. Two of the machines described have each sold to well over one hundred customers (see Chapters 8 and 9). Two more are in the final stages of product launch (see Chapters 8 and 11). The rest of the machines described in this book are research prototypes. Overall, they provide a survey of the latest trends in performance-enhancing architectures for smart information systems. The book will thus appeal to all engaged in the design or use of high-performance (parallel) architectures for nonnumeric applications. The machines featured in this book have been designed to support various information systems, ranging from relational databases to semantic networks and other artificial-intelligence paradigms. Knowledge-representation formalisms, as such, are not described in any detail. Although expertise in knowledge representations is not required for an understanding of the hardware architectures in this book, readers wishing for background information on this subject may refer to Frost. The topic of database and knowledge-base machines is a large one, and no single volume can hope to cover the field in a truly comprehensive manner. Particularly, this book does not touch on distributed or federated databases, 2 heterogeneous databases^ or related issues. However, many of the projects we have chosen to emphasize contain generic architectural ideas that may be used to support several sets of higher-level requirements. In other words, some of the machines are based on hardware designs that strive to be semantics-free. For completeness, we list the major review books, conference proceedings, and survey papers that have appeared within the last ten years on the hardware and software of database and knowledge-base …