Distributed Spatial Data Warehouse Indexed with Virtual Memory Aggregation Tree.
Marcin Gorawski, Rafał Malczok · 2004
In this paper we present a distributed spatial data warehouse system designed for storing and analyzing a wide range of spatial data. The data is generated by media meters working in a radio based measurement system. The distributed system is based on a new model called the cascaded star model. In order to provide satisfactory interactivity for our system, we used distributed computing supported by a special indexing structure called an aggregation tree. We provide a detailed description of distributed system components and their co-operation. The indexing structure operation is tightly integrated with the spatial character of the data and the data model. Thanks to a memory managing mechanism the system is very flexible in the field of aggregates accuracy. A final selective materialization of indexing structure fragments strongly increases the system’s efficiency. Basing on the tests results, we compare the efficiencies of the distributed and a singlemachine systems, and demonstrate the importance of indexing structure materialization.