Definition
A GIS (Geographical Information System) is defined on the world wide web (www.gis.com) as being
GIS combines layers of information about a place to give you a better understanding of that place
But even after reading this several times, I still have no idea what a GIS is. An alternative definition of a GIS is
a combination of information (both spatial and attribute) and functionality collected in a computer system so that a user can perform analysis
Which does not make much more sense, but at least is a starting point so that it is possible to build a system.
Components of a GIS
As a GIS is combination of information and functionality
it is obvious that it is more than just an application.
In order to deploy a complete GIS it must consist of
the information necessary to support the system as well
as the functinality, both commercial and customised.
For enterprise scale systems the most common solution is
to store the informtion in a database, but there may also
be information in files.
The functionality may come from a
vendor (such as ESRI) or be custom developed to meet the
particular requirements of a user. Together, all these
make up the GIS.
User interaction with the system is almost exclusively through functionality provided by the application, but the application will then use either GIS Files or GIS Databases as appropriate.
Technical Deployment
Current technology dictates that the functionality a GIS needs to run can be either installed on a workstation or a server and the data necessary to support the analysis can either use files or a database for storage.
Ian has considerable experience of working as the Architect for an Enterprise GIS at a large independent oil company: the GIS is deployed to over 100 users in head office as well as remote offices. Several different technologies are used to deploy the GIS to the remote offices, including web services using ArcGIS, Citrix (aka Windows Terminal Server) and desktop installation.
GIS Database
The GIS database consists of over 50 million rows of vector data (~40GB of data), ~200GB of raster data as well as interfaces to 4 other corporate databases, 5 application databases and 4 external databases (both vector and raster data). Yeah, you could say the database is a decent size.
More importantly than the size of the database is the fact that it has proven to be stable, scalable (more data is being added all the time) and able to meet just about every business requirement that has been thrown at it.