Abstract
In the past two decades, a considerable amount of research exists which uses hardware, firmware, and novel architectures to achieve the needed efficiency in implementing database management functions. However, most of the past efforts have been directed toward developing database computers for supporting a relatively primitive data model, namely, the relational model. This paper presents the design and evaluation of an Object Flow Computer (OFC). OFC is designed to efficiently support the processing of object-oriented databases. OFC employs a vertically fragmented data storage structure and a two-phase parallel query processing strategy. A set of primitive operators is defined for OFC. Depending on the performance requirement, these operators can be implemented in software running on general-purpose processors or as functions in special-purpose coprocessors. A high-level database request can be decomposed into these primitive operators and executed in parallel. OFC combines a number of known database processing techniques such as query decomposition, pipelining mode of data processing, and data flow control strategy. The performance evaluation of the proposed two-phase query processing strategy and a comparison with the conventional query processing strategy are also presented.
Original language | English |
---|---|
Pages (from-to) | 298-314 |
Number of pages | 17 |
Journal | Journal of Parallel and Distributed Computing |
Volume | 17 |
Issue number | 4 |
DOIs | |
Publication status | Published - 1993 Apr |
All Science Journal Classification (ASJC) codes
- Software
- Theoretical Computer Science
- Hardware and Architecture
- Computer Networks and Communications
- Artificial Intelligence