A general solution of the n-dimensional B-tree problem
SIGMOD '95 Proceedings of the 1995 ACM SIGMOD international conference on Management of data
Advances in the Design of the BANG File
FOFO '89 Proceedings of the 3rd International Conference on Foundations of Data Organization and Algorithms
Performance Comparison of Point and Spatial Access Methods
SSD '89 Proceedings of the First Symposium on Design and Implementation of Large Spatial Databases
A Well-Behaved File Structure for the Storage of Spatial Objects
SSD '89 Proceedings of the First Symposium on Design and Implementation of Large Spatial Databases
On the Complexity of BV-tree Updates
CDB '97 Second International Workshop on Constraint Database Systems, Constraint Databases and Their Applications
Grow and Post Index Trees: Roles, Techniques and Future Potential
SSD '91 Proceedings of the Second International Symposium on Advances in Spatial Databases
Hi-index | 0.00 |
A current trend in database architecture is to provide 'data blades' or 'data cartridges' as 'plug-in' indexing methods to support new data types. The research project which gave rise to this paper aims to test the practicality of a diametrically opposite approach: the development of a new, generic indexing technology i.e. a single indexing technique capable of supporting a wide range of data types. We believe that BANG indexing [Fre87] is now a viable candidate for such a technology, as a result of a series of extensions and refinements, and fundamental improvements in worst-case characteristics made possible by recent theoretical advances EFre95, Fre97f. The task i s therefore to test whether this single generalized technique can match the performance of several other specialized methods. This paper is devoted to the indexing of spatial extents. It describes a simple refinement of an earlier approach to spatial extent indexing based on a dud BANG representation, and compares its performance with that of the R*-tree. The results are surprising. In essence, they show that BANG indexing is able to match - and in many cases significantly surpass - the query performance of the R*-tree without incurring the heavy index optimization costs of the R*-tree. This leads to dramatic improvements in indexing times.