| New Database Engine Technology: Democratizing Data
New breakthroughs are allowing GIS users to store and manipulate all variations of geographic features within an RDBMS, while sharing the data over large networks.
By Lisa Backman Enterprisewide GIS. It's the hot trend in the GIS industry-implementing GIS across an organization so that everyone from technical GIS professionals to nontechnical managers can benefit from the analytical capabilities of the technology. GIS software now ranges from low-end desktop mapping products for casual users to desktop GIS for more technical users to high-end GIS for specialists, and runs on most hardware platforms, from PCs and Apple Macintoshes to workstations.
With GIS technology deployed across an organization, effective data management and sharing is critical. Sharing data among departments in an organization eliminates redundancy, which reduces the cost of resources (staff and technology) needed to maintain the same data and frees those resources for more important tasks.
GIS software has traditionally relied on proprietary spatial data storage formats. Attribute data are often stored in standard relational database management systems (RDBMS), which have not been configured to handle spatial data like points, lines, and polygons. And some needed data are stored with the technology first used to create it, such as CAD or spreadsheets. To accommodate users' needs to access disparate types of data and to integrate spatial and tabular data, GIS software developers have used a system of translators that allow the user to import and export different data types. This has been an effective system, but not the most efficient.
It also does not address the need for a central data repository. Trying to use GIS technology by accessing several databases is like trying to teach standard material in a class where each student has a different edition of a textbook. Sooner or later someone is going to get bad or outdated information, or they may not have access to necessary information at all.
Storing spatial information in a central database has been problematic, however, and storing spatial data with the accompanying attribute data even more so. The ideal solution would be to leverage existing RDBMS technology by being able to store spatial and attribute data together. This would allow users to access spatial information from a central location, use it for GIS as well as other applications, and let information technology (IT) managers use the data management capabilities of an RDBMS to handle spatial data. New Technology
Recognizing this need, RDBMS vendors are enhancing their technology to handle spatial data, and GIS software vendors have developed new database engine technology that allows GIS users to store spatial data in RDBMSs, such as Oracle, Informix, DB2, and SQL Server.
The Spatial Database Engine (SDE) from Environmental Systems Research Institue Inc. (ESRI), in Redlands, Calif., for example, allows the user to store and manipulate all variations of geographic features within an RDBMS, while sharing the data in a client/server environment over large networks. "This allows data to be stored in its simplest, unstructured form, which in turn allows you to operate a GIS regardless of the data model used," says Keith Ryden, ESRI's development manager for SDE. "Less data structure, better software, and faster hardware provide maximum flexibility to GIS end users."
SDE is designed to allow multiple users access very large datasets simultaneously with no degradation in performance. Query response time is measured in tenths of a second, even with dozens or hundreds of users accessing the database at once. "The foremost design goal was to provide exceptionally fast performance on large databases with multiple users," says Jack Dangermond, ESRI president.
This type of response led the hydrographic departments of Sweden and Finland to use SDE as the backbone for a joint hydrographic information system. The amount of data to be stored in this system is staggering. The Finnish Hydrographic Department alone manages data for 600 million sounding points used to measure the depth of the ocean-and the number is increasing continuously. About 100 million selected sounding points, reduced from the original survey data, will be managed in the HIS. "Fast access and handling of spatial data was the main reason to have SDE," said Pentti Junni, HIS project manager, Finnish Hydrographic Department.
The SDE-based system will be used for navigation so it is critical that data can be accessed quickly. "Safety of navigation was the main goal," said Christer Bremer, HIS project manager, Swedish Hydrographic Department. "The information managed in the HIS is vital because of the difficult navigation conditions in our region." Data for All
At the heart of this new technology is the concept of democratizing data-making spatial data available to all users, regardless of its original source, and allowing them to use it in whatever application they need.
Organizations that would benefit from using SDE technology include those that have large databases, many potential users, distributed processing needs-strongly defined data-serving functions supporting desktop applications-integrated multiple operating systems, and mixed local area and wide area network computing environments. This technology will enable all kinds of possibilities for building applications and managing spatial data cost-effectively. A New Approach
SDE is a development tool rather than a consumer software package. It is server technology and does not have its own interface-interfaces are provided by the various programs used to access the data. It uses an open application programming interface (API), which allows users to access stored data via any software package. SDE uses SQL and other technologies to access, reformat, and deliver data in the format required by the software being used.
To achieve speed while providing multiple access, the SDE provides a client/server architecture with cooperative geoprocessing. The server only responds to requests (query and retrieval), while the client performs the geometric processing (overlays, buffers, etc.). This minimizes the computational work on the server and allows many users to access the server simultaneously.
SDE supports a variety of feature types including points, lines, and polygons. While primarily a vector engine, it is well suited for serving raster images. Australia's Telstra, a telecommunications company, uses SDE and SalSoft SDE Support Library from Salamanca Software (Sandy Bay, Tasmania, Australia) to provide White Pages and Yellow Pages directory services on the Internet. Users enter the name of a business or government office. The request is passed to SDE and the Support Library. SDE locates the queried address on a set of congruent map tiles, which the Support Library then combines to build a new map centered on the address. The location is marked with an icon.
"Traditional GIS methodologies were of no use to us here," says Nick Gee, on-line mapping project manager at Telstra. "We're trying to provide high-speed access to what could be thousands of users at once, all looking at different areas of the spatial database-so we had to change our whole approach. In the end we used ESRI's SDE technology and combined it with our advanced geocoding techniques," he says. World Wide Web
With its emphasis on speed and handling multiple users, SDE technology is ideally suited for serving information up on the World Wide Web. SDE and JAVA technology are being used to share data via the web among all parties collaborating on the Humid Tropical Forest Inventory Project (HTFIP), a NASA Pathfinder Project based at University of New Hampshire. HTFIP is dedicated to making Landsat data for the Amazon, Central Africa, and Southeast Asia publicly available and to analyzing these data to quantify rates of deforestation around the globe. By the end of 1996, the HTFIP will have acquired 2,500-3,000 Landsat images and will have digested the metadata for nearly all Landsat and SPOT images.
At the university's web site users will be able to query this dataset and view the accompanying imagery. The goal is to allow users to perform sophisticated geographic queries, such as finding all Landsat imagery that intersects a particular polygon, viewing images for all cities situated within 20 km of a specified river, or finding all images from 1994 that are 90 percent cloud free and that intersect the junction of two specified rivers.
A JAVA front-end will serve as the interface to the system. The university will house the Java Server, SDE client, SDE Server, and Oracle DBMS, while users will be able to query data using a JAVA-compliant browser, such as Netscape Navigator.
SDE is letting organizations develop true enterprisewide GIS. "SDE truly spatially enables the organization," says Dangermond. "It moves spatial data management away from the traditional 'GIS-centric' viewpoint to the 'information-centric' paradigm." Storing spatial data in a central database with its attribute information ensures that everyone in the organization is working with the same, most up-to-date information. The benefits to organizations in terms of money saved, productivity gains, and return on investment are potentially enormous. About the Author:
Lisa Backman is a writer and editor with ESRI. She may be reached at 909-793-2853. Back |
