RectNet - A Distributed Geometrical Data Structure

The collection of and search for location information is a core component in many pervasive and mobile computing applications. Instead of using a centralized service for managing distributed dynamic location data, we previously introduced the concept of a distributed data structure, the socalled distributed space partitioning tree (DSPT). A DSPT is a general use distributed data structure, similar to distributed hash tables (DHTs), that allows publishing, updating of, and searching for geometrical objects. The problem of range queries on a set of points in a distrubuted scenatio space has been well studied. DSPTs generalize this problem, by allowing the keys of the objects and queries to have a spatial extension with arbitrary boundaries. In this paper we describe RectNet, a first implementation of a DSPT. RectNet is based on an binary space partitioning and torus topology. We provide an overview of RectNets architecture, algorithms, and a brief evaluation.