Comparação do desempenho dos índices r-tree, grades fixas, e curvas de Hilbert para consultas espaciais em bancos de dados geográficos / Comparison of the performance of R-tree grids files and Hilbert space filling curves performance for spatial queries on geographical databases

AUTOR(ES)
DATA DE PUBLICAÇÃO

2008

RESUMO

The development of spatial databases has been influenced by the research on spatial indexes. There has been a considerable amount of research on multidimensional indexes such as R-trees and its variants. The performance improvements when using spatial indexes for spatial operations led database developers to include R-tree support on their products. Despite these advances, there are database management systems (DBMS) that do not provide spatial indexing in their internal core. Users of this and similar DBMS may want to develop spatial applications, but may be deterred because of performance drawbacks. One obvious solution would be to implement spatial indexes in the DBMS kernel. However, such solution may not be feasible, either because of lack of access to the source code or due to time and cost constraints. This work investigates an alternative way of including spatial indexes in DBMS. We consider indexes such as Hilbert space-filling curves and Fixed Grid developed on top of TerraLib GIS Library. Our main goal is to provide, through this library, the functionalities that are not available in a DBMS that do not include spatial index support and verify the performance between the indexes and the native mechanisms of the SGBDs like B-Trees and R-Trees.

ASSUNTO(S)

desempenho de sistemas computacionais distribuição espacial computação aplicada computer science índices indexes computer systems performance efficiency espaço de hilbert spatial distribution eficiência hilbert space

ACESSO AO ARTIGO

Documentos Relacionados