STAR-Tree: An Index Structure for Efficient Evaluation of Spatiotemporal Authorizations

Typically, geospatial data include satellite imagery, digital orthophoto quads, maps, digital line graphs, census data, voter registration, land ownership data, and land use data. This data is considered sensitive based on its location (characterized by its longitude and latitude), resolution, and the time of capture, among others. Since both objects and authorizations are associated with spatial and temporal attributes, in order to process access requests efficiently, it is essential that they both be accessed using some sort of index structures. However, processing an access request under this approach requires searching two indexes - one the object index and the other the authorization index. In this paper, we propose a novel index called STAR-Tree, a Spatio Temporal Authorization-driven R -Tree, that can uniformly index both spatiotemporal objects and the authorizations that govern access to them. STAR-Tree is an extension of R-tree that allows objects of different resolutions be indexed based on their spatial and temporal attributes, as well as allows layering of spatiotemporal authorizations on the tree itself. Compared to the previously proposed RMX-Quadtree, STAR-Tree enjoys several advantages. First, the 3 dimensional nature of the STAR-Tree accommodates the temporal dimension. Second, the STAR-Tree imposes no restrictions on the region covered by the geospatial objects. Third, in the STAR-Tree images of the same resolution may overlap with one another. We demonstrate how such a tree can be constructed and maintained, and show how access requests can be processed in an efficient manner.