The concepts and capabilities of a database computer (DBC) are given in this paper. The proposed design overcomes many of the traditional problems of database system software and is one of the first to describe a complete data-secure computer capable of handling large databases. This paper begins by characterizing the major problems facing today's database system designers. These problems are intrinsically related to the nature of conventional hardware and can only be solved by introducing new architectural concepts. Several such concepts are brought to bear in the later sections of this paper. These architectural principles have a major impact upon the design of the system and so they are discussed in some detail. A key aspect of these principles is that they can be implemented with near-term technology. The rest of the paper is devoted to the functional characteristics and the theory of operation of the DBC. The theory of operation is based on a series of abstract models of the components and data structures employed by the DBC. These models are used to illustrate how the DBC performs access operations, manages data structures and security specifications, and enforces security requirements. Short Algol-like algorithms are used to show how these operations are carried out. This part of the paper concludes with a high-level description of the DBC organization. The actual details of the DBC hardware are quite involved and so their presentation is not the subject of this paper. A sample database is included in the Appendix to illustrate the working of the security and clustering mechanisms of the DBC.

The concept of the data utility as a database system capable of supporting a data model, a large-scale on-line store, and concurrent user access has emerged in recent years. New technology can make the notion of database computers—a specialized hardware system for database use—a viable one. A summary of previous activities and a discussion of the state of the art of database computers is given. An attempt is made to suggest guidelines for future database computer architecture research.

An algorithm is given for synthesizing a computer program from a trace of its behavior. Since the algorithm involves a search, the length of time required to do the synthesis of nontrivial programs can be quite large. Techniques are given for preprocessing the trace information to reduce enumeration, for pruning the search using a failure memory technique, and for utilizing multiple traces to the best advantage. The results of numerous tests are given to demonstrate the value of the techniques.