Home Articles Enterprise Geographic Information Servers

Enterprise Geographic Information Servers

Dr David J Maguire
Dr David J Maguire
Director of Products, ESRI
[email protected]

The earliest systems were built to run on mainframe and mini-computers, then came workstations, next PCs, and now system designers are recommending web-based technologies for distributed systems. Enterprise Geographic Information Servers are a new type of GIS architecture that supports access to potentially thousands of distributed users from a centralized location. The industry standards-based approach to data management, application development, and data and processing access makes them ideal for providing GIS services to a wide range of distributed users.

Throughout the course of the last forty years, Geographic Information Systems have been created using a variety of contemporary computer architectures (Coleman 1999). The earliest systems were built to run on mainframe and mini-computers, then came workstations, next PCs, and now system designers are recommending web-based technologies for distributed systems (Peng and Tsou 2003). A new architecture is currently emerging based on some new technologies that allows all GIS functions to run in a centralized server environment and be accessed from any device on a network. This paper will attempt to describe this new Enterprise Geographic Information Server approach and its implications for building truly distributed enterprise GIS.

GIS comprise three major subsystems that are concerned with mapping, spatial analysis (geoprocessing) and data management. A useful general purpose GIS needs capabilities in all these areas.

The ‘mapping subsystem’ deals with the manipulation and visualization of geographic information. It is here that users will find projection and datum transformations, map to page transformations, symbology models, map displays, 2D and 3D visualization tools, editing components, and the framework for interacting with geographic data in the form of a map. The ‘spatial analysis subsystem’ implements a series of geographic analysis functions for the two classic geographic operations of proximity and overlay analysis, as well as functions for data conversion (import CAD data, export graphics, etc.), grid analysis (watershed modeling, intervisibility analysis, etc.), and 3D analysis (calculation of slope, aspect, etc.).

In the past few years it has become common practice to use commercial off the shelf relational database management system (RDBMS) products – such as Microsoft Access, IBM DB2, Oracle and Microsoft SQL Server – to store and manage geographic information (Longley et al 2001). RDBMS have been widely used in GIS because they allow users to create a single, centralized data repository (avoiding redundancy and duplication), they facilitate data sharing (by establishing de facto standards), they support multi-user editing of continuous geographic databases, and they allow users to employ RDBMS backup and recovery tools. However, the limited support for advanced geographic data types (Shekar and Chawla 2003), the weaknesses of SQL as a geographic data access and programming language (Egenhofer 1982), and concerns about scalability for operations such topology management (Hoel et al 2003) have instigated a re-examination of the role of RDBMS in GIS and have prompted the development of Enterprise Geographic Information Servers.

Fig 1 The three subsystems that are needed for a well-rounded GIS platform