Graph database retrieval based on metric-trees

Abstract

M-trees are well-know structures used to speed-up queries in databases. In this paper, we evaluate the applicability of m-trees to graph databases. In classical schemes based on metric-trees, the routing information kept in a metric-tree node is a selected element from the sub-cluster that represents. Nevertheless, defining a graph that represents a set of graphs is not a trivial task. We evaluate different graphs-class prototype as routing nodes in the metric tree. The considered prototypes are: Median Graphs, Closure Graphs, First-Order Random Graphs, Function-Described Graphs and Second-Order Random Graphs. © 2012 Springer-Verlag Berlin Heidelberg.