EXERCISES AND EXAMINATIONS OF SOFTWARE ENGINEERING TECHNIQUES FOR THE IMPLEMENTATION OF LARGE-SCALE DATABASE SYSTEMS: THE RESULTS OF A MULTI-BACKEND DATABASE SYSTEM IMPLEMENTATION.

摘要

In this report we propose a methodology, database systems implementation methodology (DBSIM), for the development of large-scale database systems. The methodology and the results of using the methodology in the implementation of a multi-backend database system are presented. The methodology includes software engineering techniques for managing the database systems effort and for developing the database systems. The management techniques include a development strategy, a team organization and techniques for software quality assurance. The software development techniques include techniques for the design, coding and testing phases of the software life cycle.;A review of different solutions to the three categories of issues and our solutions to these issues are described. The database systems implementation methodology (DBSIM) is then developed. The techniques used in DBSIM and the rationale for choosing them are described in detail. An application of the DBSIM to development of a large-scale database system, i.e., the multi-backend database system (MDBS), is reported. Our database systems engineering experience is also presented.;The contributions of the work to database systems engineering, software engineering and database systems are discussed. The directions for further research and experiment in these areas are presented.;The issues that must be addressed for a database system development effort are presented. We categorize the issues as "management issues", "system hardware and software issues" and "database system implementation issues." The management issues include the choice of an implementation strategy, the organization of implementation personnel, the training and expertise of implementors, the turnover of programming teams and the assurance of software quality. The system hardware and software issued include the selection of hardware, the choice of a systems programming language, the choice of operating systems and the techniques for software development. The database system implementation issues include the interface with operating systems, the message-passing mechanisms, hardware and software requirements, the level of I/O operations, the organization of directory data and user data, the concurrency control mechanisms and the supporting software.