The Whole Enchilada

Enterprise GIS servers could be an easier, more economical way to distribute full GIS functionality to many desktops.

by / March 28, 2005
Traditional Web-based GIS tools allow governments to publish interactive mapping applications on a Web site, allowing citizens to use the maps for their own purposes while simultaneously providing a way for agencies and jurisdictions to share geographic data.

Web-based systems are popular because users can access them from any computer with an Internet browser. They are also easily and inexpensively maintained since all the software resides on a central server -- IT staff needn't visit users' machines to troubleshoot software problems or install upgrades.

Though the public sector uses Web-based GIS in many ways, the approach is not without drawbacks.

One problem is that Web-based GIS applications aren't as powerful or versatile as the systems GIS professionals run on their desktops.

Historically software developers took a small number of desktop GIS functions -- mapping, geocoding and point-to-point routing -- and turned them into an Internet service designed with an Internet-based architecture, said David Maguire, director of products and international at GIS developer ESRI in Redlands, Calif.

"It fit nicely into the Internet, but it didn't have all the functional capabilities high-end GIS users had come to expect and require," Maguire said.

Enterprise GIS Servers
Maguire said ESRI saw a market for an alternative to both desktop-based GIS solutions and the less-functional, Web-based GIS applications created with tools such as its own ArcIMS.

That alternative is what ESRI calls an enterprise geographic information server (EGIS). ESRI debuted its first EGIS product, ArcGIS Server, last year. As with any new product offering, it's too early to tell whether the EGIS product line (or other competing efforts by GIS vendors) will gain a foothold in the government marketplace.

One thing is clear: Many governments and government agencies realize the value of pursuing an enterprise GIS architecture. In part, it allows governments to make better use of GIS data and investments into hardware and GIS software.

Perhaps more importantly, an enterprise GIS architecture could create an environment more conducive to information sharing.

GIS Utility
Oregon's Geospatial Enterprise Office (GEO) is pursuing an enterprise GIS architecture to improve information sharing, and plans to implement ArcGIS server to support a "GIS utility" the state is currently developing.

The utility project is in its early stages, said Cy Smith, Oregon's statewide GIS coordinator. Under the first portion of the three-part development program, Oregon is assembling data sets -- for use by state, local, federal and tribal agencies in Oregon -- covering 14 "themes," such as transportation, hydrography, elevation, tax lots and administrative boundaries.

The second component is creating a central architecture allowing the state to deliver GIS data and functions over the Web. Oregon already uses ESRI's ArcIMS and ArcSDE products with an Oracle database manager to distribute data and maps via a Web server.

"We need a central enterprise server that has some tools that will allow us to make the data more accessible in real time," Smith said. "Right now, our ArcIMS application is very simple. You can view the data. You can zoom around and do some simple queries, but you're not really making full use of the data -- as you would in desktop GIS software with full-blown functionality."

For the third component of the program, Oregon will create an engine allowing GIS applications to draw data from multiple sources -- some of it housed in a central server, and some housed in servers owned by agencies that need to maintain their own data.

Data that falls under the transportation theme -- information on street center lines, address ranges, highway attributes, etc. -- might reside with any number of agencies at different levels of government, Smith said, and Oregon's data clearing-house combined with the GIS utility will allow an application to seamlessly combine data drawn from different locations.

"If you're going to route an emergency vehicle properly, you've got to have all of those available in one data set in one location, or at least made to look transparent as if it's one data set," he said.

Extra Horsepower
Under an enterprise GIS server architecture, all aspects of the GIS, including mapping, spatial analysis and database management, run on a central server. Users working on thin clients -- PCs or other devices equipped with browsers -- use a local- or wide-area network to access those GIS applications.

Staff needing more sophisticated functions can access the system from a "thick" client machine, which runs some of the necessary software locally. The architecture also supports handheld systems that interact with the server while connected to the network, but that also function while disconnected.

ArcGIS Server includes full geographic analysis functions, high-end cartography and other capabilities, uses standards-based application programming interfaces and supports extensible markup language/simple object access protocol (XML/SOAP), Java and .NET, Maguire said.

The EGIS approach promises many advantages of other Web-based systems -- scalability, easy access and inexpensive maintenance -- while delivering full GIS functionality because, according to ESRI, instead of trying to adapt a desktop system to the Web environment, the ArcGIS Server was designed for the Internet.

Not Quite Yet
The Kansas Data Access and Support Center (DASC), like its Oregon counterpart, provides a GIS clearing-house for state and local agencies. Unlike Oregon's GEO, the DASC currently has no plans to procure and deploy ArcGIS Server.

"I don't think many people are, because it's pretty new," said Kenneth Nelson, GIS coordinator at the Kansas Geological Survey and manager of the DASC, while acknowledging that future customer requests might push the DASC to consider an enterprise GIS server in the future.

The DASC, housed at the University of Kansas in Lawrence, currently uses ArcIMS to distribute interactive maps and basic GIS applications to users via the Web. Nelson said ArcIMS is an easy and cost-effective way to allow people who aren't trained professionals to use the data, which was previously only useful to GIS professionals with the expertise to locate the data they wanted, and download, unzip and import it into the applications they were running on the desktop.

An enterprise GIS server would allow the state to deliver all functions currently available in the ArcGIS family of products over the Internet, Nelson said.

While the enterprise GIS server approach offers a major advantage, Kansas will not deploy it in the near future because DASC staff simply have too many other projects on their to-do lists.

"The investment in time to get up and running and develop applications in that environment is pretty large," Nelson said, because ArcGIS Server is more expensive than ArcIMS, and applications developers must climb a steep learning curve before they can use it.

So far, DASC staff members have developed all the applications that clearing-house users have requested. But as more users at state and local agencies come to understand the power of geographic data, it's very likely they will start asking for applications that strain the limits of ArcIMS, Nelson said.

"I think as their needs mature, that will probably push us toward that server technology, where they're trying to do more sophisticated query or analysis or more sophisticated data development."
Merrill Douglas Contributing Writer
Platforms & Programs