The Dynamic-BrickR access method for mobile objects

The database storage and organization of data describing the evolution of mobile objects is an open challenge. Data must be managed in efficient structures with respect to both the storage space consumed and the data access through the means of these structures. We introduce in this paper a new indexing method that improves the data access and, consequently, the queries processing efficiency. The Dynamic-BrickR access method uses two structures: an underlying permanent R*-Tree structure, and an in-memory dynamic space grid structure, that it used for building the terminal nodes to feed the R*-Tree. Experiments show significant improvements of the Dynamic-BrickR method over the R*- Tree index, regarding the dead space and the overlapping area. The Dynamic-BrickR can be used in answering spatial, temporal and spatio-temporal queries. Between them, spatial and spatio-temporal queries benefit by the improvements the Dynamic-BrickR brings at the space structuring level and are answered by fewer data access operations.