TITLE: | Spatial Database Management Systems: A Tour |
PRESENTER: | Shashi Shekhar : Biography , Homepage |
AFFILIATION: | Computer Science Department, University of Minnesota. |
URL: | http://www.cs.umn.edu/~shekhar |
VIDEOS and SLIDES:
SLIDES:
Classical relational DBMS (RDBMS) often perform poorly when applied to spatial data sets because of the following reasons. First, RDBMS conceptual data models (e.g. entity relationship diagram(ERD) or UML) document all relationships explicitly assume a sparse set of relationship among entities. However, spatial entities have a rich set (almost completely connected) of implicit relationships (e.g. distance, direction) leading to a cluttered conceptual model. Second, RDBMS query languages (e.g. SQL) have a limited set of data-types (e.g. numbers, text-strings) leading to a large semantic gap with the needs of spatial querying (e.g. Google Maps, GPS devices). Finally, RDBMS indexes (e.g. B-tree, hashing), and query optimizers are designed for linear-ordered one-dimensional values such as numbers, however, spatial data is embedded in a multi-dimensional space.
Thus, SDBMS (e.g. ESRI GeoDatabase/SDE, Oracle SDO, DB2 SDC, Postgres postgis) have emerged over last decade to meet the unique needs of modeling, querying, and analysis of very large spatial datasets. We uncover SDBMS at three levels, namely, conceptual (e.g. Pictogram enhanced ERDs), logical (SQL3/OGIS query language) and physical (e.g. R-tree index). Trends (e.g. spatial networks and spatial data mining) are discussed briefly.
KEYWORDS: Spatial Datasets, Spatial Databases, Relational Databases.
ACKNOWLEDGMENTS: This work was supported in part by the National Science Foundation, the U.S. Department of Defense, the National Aeronautics and Space Administration the Federal Highway Authority, and the University of Minnesota (e.g. Center for Transportation Studies).
NOTE: Some of the results discussed in this talk appeared in the following publications: