Towards historical R-trees

Publisher Website
Google Scholar
Abstract
R-trees are the "de facto" reference structures for index- ing spatial data, namely the minumum bounding rectangles (MBRs) of spatial objects. However, R-trees, as cur- rently known, do not support the evohition of such MBRs. Whenever an MBR evolves, its new version replaces the old one, which is therefore lost. Thus, an R-tree always shows the current state of the data set, not allowing the user to query the spatial database with respect to past states. In this paper we extend the R-tree in such a way that old states are preserved, allowing the user to query them. The proposed approach does not duplicate nodes which were not modified, thus saving considerable storage space. On the other hand, the query processing time does not depend on the number of past states stored.

This publication has 0 references indexed in Scilit: