Multidimensional Access Methods: Trees Have Grown Everywhere
Timos Sellis, Nick Roussopoulos, Christos Faloutsos · Very Large Data Bases · 1997
This paper is a survey of work and issues on multidimensional search trees. We provide a classification of such methods, we describe the related algorithms, we present performance analysis efforts, and finally outline future research directions. Multi-dimensional search trees and Spatial Access Methods, in general, are designed to handle spatial objects, like points, line segments, polygons, polyhedra etc. The goal is to support spatial queries, such as nearest neighbors queries (find all cities within 10 miles from Washington D.C.), or range queries (find all the lakes on earth, within 30 and 40 degrees of latitude), and so on. The applications are numerous, including traditional database multi-attribute indexing, Geographic Information Systems and spatial database systems, and indexing multimedia databases by content. $‘rom the spatial databases viewpoint we can dist,inguish between two major classes of access methods: