Parallel computer architectures for relational databases
Thesis/Dissertation
·
OSTI ID:6277895
The main purpose of this dissertation is to explore advanced computing techniques for the field of information processing. A number of existing database machines are evaluated according to their district computer architectures. Two new database machines are proposed: (1) IVH database machine: a data structure (called IVH) is implemented in a multiprocessor environment. IVH is evaluated by comparing with n-ary and inverted file systems. This system is developed for efficiently implementing relational operations, handling complex data models and supporting distributed database systems. (2) BIFRAM database machine: the topology of parallel sorting network is studied, and several processing algorithms are established for performing all computational intensive relational primitives in the network. A two-dimensional processor array (ROPU) is designed to implement these algorithms using a specific control mechanism. ROPU is utilized for relational algebra machine design and compared with other VLSI hardwares. Further BIFRAM relational database machine is developed for Fifth Generation Computer Systems. This machine has several characteristics: one major processing unit, no extra step for multi-relations, small main memory size, hardware similarity, well distribution function and great computing power.
- Research Organization:
- Wayne State Univ., Detroit, MI (USA)
- OSTI ID:
- 6277895
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallel processing of relational algebra operations
Architectural alternative for data base
Pumps architecture for pattern analysis and image database management
Thesis/Dissertation
·
Wed Dec 31 23:00:00 EST 1986
·
OSTI ID:5924486
Architectural alternative for data base
Thesis/Dissertation
·
Mon Dec 31 23:00:00 EST 1984
·
OSTI ID:5127458
Pumps architecture for pattern analysis and image database management
Journal Article
·
Fri Oct 01 00:00:00 EDT 1982
· IEEE Trans. Comput.; (United States)
·
OSTI ID:5303495