Distributed space partitioning trees (DPSTs) (Heute lbeck and Hemmje, 2006) describe a distributed spatial index for storage and access of the so-called location knowledge. Location knowledge is acquired in a complex distributed process through positionin g, tracking, and program logic to support location- based services which are based upon knowledge about the physical location, shape, and size of real and virtual entities, such as persons, vehicles, cities or other location-based services. Location knowled ge is needed at diverse applications in geographical info rmation systems (GIS), virtual collaboration system s and mobility management. The previous implementatio n of a DSPT, RectNet (Heutelbeck and Hemmje, 2006), was built on an adaptive binary tree shaped network topology. The use of a recursive space partitioning can be useful for load balancing. Howe ver, previous implementation requires complex operations to restructure the network. These operat ions cause significant bursts in network traffic. I n this paper we present VoroDSPT, a new DSPT implementatio n, by using the geometric structure Voronoi Diagram as a topology that provides both greedy-rou ting for supporting efficient spatial queries and information sharing, providing the base for implementing future efficient heuristics of load balancing.