Spatial Data Architecture

Why a Spatial Data Architecture?

Consider the construction of a building. There will be plenty of different types of materials from which the construction can take place, such as bricks, wood, tin, tiles, nails, paint, and so on, but these need to be put together in some manner. On their own, they are useless, but together (as a building), they are very useful. The building will have a particular style and structure, which is referred to as its architecture. This architecture is used in the process of designing and constructing the building. Of course, not all the detail of a building is defined in its architecture, since there may be many different buildings all utilizing the same architecture. However, the architecture defines the framework in which the building structure will be detailed.

In a similar manner, spatial data need to be organized within an architecture. Spatial data comprise a range of data types. When building a GIS, data of many different [Page 401]types need to be put together within an integrated structure to make them useful. An SDA defines the structure or framework in which spatial data will be represented in a GIS. Note that an SDA does not refer to the data itself, nor does it necessarily need to provide all the detail for storage, unlike a database management system. Rather, it identifies the framework used to further design and implement a GIS database.

Building an SDA is an important step in the process of designing a GIS. A poor architecture will lead to inconsistencies, inefficient access and retrieval, and even the inability to execute certain functionality. A good and efficient SDA is fundamental to the effective implementation, operation, and extensibility of a GIS. Hence, it is important that sufficient time and effort are spent on designing and analyzing an SDA appropriate to the tasks and applications of the GIS being implemented.

SDA Components

With reference to an SDA, the architecture is determined by the types and properties of spatial data that will be contained, as well as the environment in which the data will be stored and retrieved. An SDA will, therefore, consider the following aspects:

• What types of data are to be included
• How the data components are stored and linked together
• Where the data will be located

Spatial data may comprise a range of types, including vector data (points, lines, and polygons), raster data (cells), attributes (that are properties of the vector or raster features), images (e.g., remote sensing, aerial photographs, scanned maps), topological relationships, and metadata (information about the data). The SDA will specify what data types are included. For example, an SDA for a road management application may specify that data types include linear road segments and routes as vectors, road type and speed limit attributes, elevation raster data, and aerial imagery. Furthermore, an SDA must be able to accommodate large volumes of data with representations at multiple scales and varying accuracies.

...