Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Parallel processing of relational algebra operations

Thesis/Dissertation ·
OSTI ID:5924486
The main motivation of this dissertation is to study the computational complexity of relational algebra operations and to develop parallel and pipelined VLSI algorithms for their efficient implementation. The complexity study is performed in terms of the number of primitive operations required to process relational algebra operations. This complexity study takes into account not only the size of the operand relations, but also the number of distinct elements in the columns of interest. The hardware algorithms are aimed towards a VLSI implementation. Consequently, regularity and uniformity are necessary in order to make the design cost-effective. A Bus connected Cellular Array (BCA) processing unit is proposed to implement relational algebra operations such as projection and set operations. The latter operations are allowed to have multirelations as operands. Both internal as well as external VLSI algorithms are developed. The idea of incorporating a BCA processing unit into a back-end relational database system is investigated. A performance analysis in terms of the system overall parameters shows that a significant performance improvement is achieved if the BCA processing unit is made part of the back-end relational database system.
Research Organization:
Wayne State Univ., Detroit, MI (USA)
OSTI ID:
5924486
Country of Publication:
United States
Language:
English