A Few new GIS and enterprise technologies
Jack Dangermond
Environmental Systems Research Institute, 380 New York Street, Redlands
CA 92373-8100 U.S.A.
Abstract
The pace of development of both communications and computer hardware and software
capabilities continues to be quite rapid, resulting in a number of new technologies for GIS
(geographic information system) and enterprise information systems. One of these is
technology for performing rapid spatial queries on information stored in standard
relational databases either within an organization or on servers scattered across a network.
Another is object technology for embedding GIS capabilities in new or existing
applications on desktop computers. A third is the ongoing development of new GIS-related
capabilities for the Internet and World Wide Web (WWW), including
componentized GIS capabilities and metadata standards. Other technical developments
are also important. All these recent technical developments are about to make profound
changes in the way in which GIS can be applied to AM/FM. By spatially enabling
enterprise information systems they will leverage AM/FM investments made over many
decades in some of the largest information warehouses in the world. They will pave the
way for AM/FM spatial information resources to flow on the Information Highway which
the Internet and the National Spatial Data Infrastructure (NSDI) will soon provide. The
implications of these changes are discussed.
Introduction
More and more organizations are recognizing that their data are closely linked to
location; as a result they are moving to apply GIS to their mission-critical operations. As
AM/FM-using organizations develop and use GISS they face various problems common
to most of today’s information technology organizations, such as managing increasingly
large databases with increasingly diverse data types, providing enterprise-wide access to
the organization’s data, meeting the needs of increasing numbers of casual users, working
in expanding clientiserver networked environments, and providing information via the
Intemet.
The capabilities described here were designed to meet some of these needs.
Processing Spatial Data in Standard Relational Databases
Many organizations have very large quantities of data stored in relational database
management systems (RDBMSS). The data structures of these RDBMSS are increasingly
being extended to provide common means of access to a wider variety of data types,
123.including spatial data. Such developments mean that spatial information can be brought
into the computing mainstream of standard, transaction-processed databases and that
organizations can combine their management information system (MIS) and their GIS in
a single relational database.
For organizations such as utilities, however, these databases are very large, often in the
range of tens to hundreds of gigabytes or more, and while the speed of transactions with
the database have been acceptable for text and numeric data, until recently most
transactions with spatial data have been much slower, often “long transactions,” requiring
minutes or much longer periods of time for execution. Fortunately, a new data processing
engine technology has become available which makes a wide variety of management and
analysis operations on spatial data about as fast as transactions on other forms of data.
This technology has been termed a spatial database engine. It operates in a standard
client/server environment, and resides on both client computers (including desktop PCs)
and server computers. The software is based on open system design principles. The
spatial server is optimized for queries (both conventional SQL and spatial queries) and
retrievals, and provides extremely fast system response times (a few seconds), even with
hundreds of simultaneous client users. This functionality was designed for AMIFM and
other large organizations in which very large RDBMSS reside on mainframe computers
which are accessed by hundreds or thousands of users.
This new technology provides standard software tools for spatial data management and a
wide range of geoprocessing tools is available to programmers and users.
Geographic features are represented in a geometric data model, complemented with
relational database data types (e.g., integer, date, BLOBS, etc.) for representing attribute
information. The data model is feature oriented: geometric data and associated attributes
are stored together and a single disk access retrieves both.
The technology is designed for use by system developers and integrators, providing them
with the tools they need to construct custom applications to meet a wide variety of
specific needs in enterprise, client/server, networked environments. It is a large set of
spatial operators together with a spatial application programming interface (API). This
technology’s functionality can be built into applications in ways which are transparent
(invisible) to the user, minimizing the need to educate users on GIS technology.
By providing familiar kinds of access to new classes of data, this kind of data processing
engine technology can expand the range of applications for corporate databases as well as
leading toward development of next-generation spatial objects and data types.
Embedded Object Technology
Among the successes of object-oriented technology has been its use in the rapid
development of custom applications and custom user interfaces through the use and reuse
by software and application developers of various kinds of interchangeable software
components. By enhancing the speed and reliability of application programming, such
124.object technology has been especially useful to AM/FM organizations engaged in rapid
prototyping and development of new applications in such areas as system design,
emergency response, and call-before-you-dig.
A recent GIS software development makes use of this approach; it is a set of object-oriented
tools which supports a building block approach to the rapid development of
mapping and related GIS applications.
A variety of components have been developed which provide such GIS capabilities as
desktop mapping, data display, handling of images and multiple map layers, SQL query,
spatial analysis, GPS connection, data publishing on a network, and the like.
