The Grid File: An Adaptable, Symmetric Multikey File Structure
ACM Transactions on Database Systems (TODS)
The interpolation-based grid file
PODS '85 Proceedings of the fourth ACM SIGACT-SIGMOD symposium on Principles of database systems
Systolic (VLSI) arrays for relational database operations
SIGMOD '80 Proceedings of the 1980 ACM SIGMOD international conference on Management of data
Interpolation-based index maintenance
PODS '83 Proceedings of the 2nd ACM SIGACT-SIGMOD symposium on Principles of database systems
Storage mappings for multidimensional linear dynamic hashing
PODS '83 Proceedings of the 2nd ACM SIGACT-SIGMOD symposium on Principles of database systems
IBM Systems Journal
| Hi-index | 0.00 |
The I/O bottleneck represents a major problem in architectures hat have been pro-posed to implement hard database operations such as join and projection. It is recognized that solutions to this problem cannot be based on new database machine architectures alone if satisfactory performance goals are to be atteined. A case in point is illustrated by the comparison of cellular/associative and in- stream pipeline based architectures. A metho- dology based on a global order preserving and dynamic partitioning is presented. The relevance of this approach to the solution of the I/O bottleneck problem is demonstrated through the efficient parallel processing of the join and projection operations. Finaly, this methodology is incorporatated into a specific database machine architecture; namely, the RAP.3 database machine. The partitioning strategy has been previously proven to be supe- rior to he other known methods.