150 likes | 274 Views
Kedar Gawande Position Paper ITK 478 Fall 2006 10/ 04 / 2006. Background for Topic Selection Success & Penetration of OO Methodologies Inherent Limitations in Relational DBMS Increasing Migration towards Object based Technologies in the relational world.
E N D
Kedar Gawande Position Paper ITK 478 Fall 2006 10/ 04 / 2006
Background for Topic Selection • Success & Penetration of OO Methodologies • Inherent Limitations in Relational DBMS • Increasing Migration towards Object based Technologies in the relational world
King County, WA - Homeland Security Public Safety Portal Development • Utah Department of Transportation (UDOT) Oracle Spatial Transportation Asset Management • Bay Area Geological Hazard Abatement Districts use GIS Mashup to Identify and Manage Critical Data
GIS ,GPS, CAD, CAM, Multimedia Information Systems, Image Processing • Data management & processing needs • Size, complexity of applications • Needs translate into requirements for a DBMS
Representation of space & elements within space • Real World Entities & its Spatial Representation • City, Street Intersection • Streets, River, Boundary • County, State, Parks, • Sample Queries • List all Emergency Exits within STV & Directions to the closest one from STV 108 • Find length of a river within a state
Consider a spatial dataset with: • County boundary (dashed white line) • Census block - name, area, population, boundary (dark line) • Water bodies (dark polygons) • Satellite Imagery (gray scale pixels) • Storage in a SDBMS table: create table census_blocks ( name string, area float, population number, boundary polygon );
Spatial DB features: • Spatial Model • Spatial ADTs ( Abstract Data types) • Spatial Operations like distance, overlap • Spatial Querying like Spatial Join • Spatial Indexing
Find Name & Area of Districts with Area more than 500 sq miles & find name and length of rivers that flow through that county. • Select C.name, C.district.name, C.district.Area(), R.name, R.length() from County C where C.district.Area() > 500 AND overlap(C, R)
User Defined Data Types – If you can see it, you can model it • User defined Behavior – Different for different types of shapes in space • Inheritance – Polygon for city can inherit from polygon for state • Additional geometric types than relational • Enhanced Querying – If you can think it, you can query it • Spatial Indexing improves performance
Why not RDBMS ? • Primitive data types – number, string etc • Cannot model spatial objects on these data types • Limited number of spatial object representations like points, lines • Difficult to model other shapes, Needs breakdown of shapes into lines and edges • Scalability • Extensibility • Complexity
Included in Enterprise Edition of Oracle 10g • Supports 2 schemas • Object Relational • Relational ( Limited support) • Storage, Retrieval, Update, and Querying • Per Geometry Instance • Single row • Single column of type MDSYS.SDO_GEOMETRY • Data for County
Every shape / spatial data type can be designed using this data type SDO_GEOMETRY CREATE TYPE sdo_geometry AS OBJECT ( SDO_GTYPE NUMBER, SDO_SRID NUMBER, SDO_POINT SDO_POINT_TYPE, SDO_ELEM_INFO MDSYS.SDO_ELEM_INFO_ARRAY, SDO_ORDINATES MDSYS.SDO_ORDINATE_ARRAY );
New age applications require complex data types and complex operations • Spatial Data models are necessary to represent them • Relational systems are inefficient to handle such data models • Object Relational systems have necessary and sufficient features to handle them
Comments, suggestions welcome • Don’t forget the evaluation