Directions Magazine (DM): If all of this software is open source and free, why should I pay OpenGeo for it?
Rolando Peñate (RP): Because there’s more to a software product than the license of the source code. While closed source software may have explicit license costs, all software (including open source) has maintenance, operating and other related costs. OpenGeo’s mission is to lower those costs while continually enhancing the functionality of open source, and our customers value that highly. We expand on this idea in our white paper, "The Value of the OpenGeo Suite."
Those who have seen Paul Ramsey’s 2011 keynote at FOSS4G Denver are familiar with Pentaho’s beekeeper model, which sets up a metaphor between software development and beekeeping. Although honey is mostly free for the taking, generally consumers are willing to pay a beekeeper to cultivate, package and distribute honey rather than dealing directly with bees themselves. While unsupported open source software shifts costs to the end user, this isn’t an issue if the end user or his/her enterprise has expertise in the relevant software and is willing to pay for support using staff time. For enterprises lacking that internal expertise, commercial open source like the OpenGeo Suite provides the option to save time and reduce direct labor costs by outsourcing maintenance and support to experts. We encourage enterprises to use unsupported open source software, but we believe that most will want an organization with expertise in developing, packaging, documenting and supporting said software when it comes time to deploy in production environments.
DM: The OpenGeo Suite includes several components. Together they form a stack for small to large enterprise GIS. What does each one do? Do users typically need to add anything (besides hardware and data) to get up and running?
Figure 1: Components of the OpenGeo Suite
RP: The OpenGeo Suite packages several open source projects into a modular enterprise spatial IT solution:
- PostGIS, a spatial database extension for the PostgreSQL database
- GeoServer, a geospatial server capable of connecting to a variety of data back-ends (e.g., Oracle Spatial, Microsoft SQL Server, Esri ArcSDE) and publishing as OGC services (e.g., WMS, WFS, WCS, WPS)
- GeoWebCache, a robust caching engine to improve the speed and performance of serving map tiles
- OpenLayers, a JavaScript library for displaying map data in Web browsers
- GeoExt, a JavaScript library for creating rich map applications in Web browsers
Given the current competitive landscape, it’s easy to see the OpenGeo Suite as an enterprise GIS alternative—yet we see the OpenGeo Suite as a flexible stack for spatial IT (see: Paul Ramsey on Spatial IT). Since there isn’t anything necessarily special about spatial data, enterprises shouldn't be forced through a specialized GIS workflow guarded by GIS analysts. Spatial data is increasingly just one aspect of the much larger information technology story within a given enterprise; our software is designed to expose and utilize this data through the Web in a similar way to non-spatial data.
The New York City Department of Information Technology and Telecommunications (NYC DoITT) is using our software alongside its existing information technology infrastructure in just this way to provide useful information to NYC residents. Rather than replace the city’s existing Oracle infrastructure with yet another monolithic enterprise solution, DoITT was able to expose its existing geospatial data to the Web using the OpenGeo Suite without sacrificing workflows that had already been adopted across NYC agencies. This enabled DoITT’s developers to not just create NYC-specific basemaps but to also develop publicly-facing Web applications that provide targeted information about street closures, snow plow status, and other street conditions as needed.
DM: The new 3.0 release "makes significant advances across the stack in terms of through-the-Web processing of geospatial data.” What exactly is “through-the-Web processing”? How is it different from other desktop or Web processing?
Figure 2: OpenGeo made this prototype for the USGS as an example of a Web-based processing tool that would allow a broader range of users to engage in editing NHD data in a way that respects the strict topology rules inherent to the NHD (Click for larger image).
RP: A major difference between desktop processing and the OpenGeo approach is that we are building our platform specifically for the Web. In contrast, traditional GIS requires “certified” experts with extensive training to pull data into expensive desktop tools and provide derivative data. The OpenGeo Suite exposes similarly powerful functionality on the Web to make advanced spatial processing available, accessible and immediate to IT professionals. Spatial analysis and visualization can then be integrated with the tools that IT professionals use every day to work with their large and time-sensitive data.
Our priority is to make powerful processing tools available on the Web for broad consumption using OGC’s Web Processing Service (WPS) specification. This not only means performing spatial processes on the server but also designating the server as the place where new processes are defined. Combining WPS and GeoScript, a Web developer can create processes that perform complex analyses using familiar scripting languages like Python or JavaScript. This enables IT professionals to build Web applications that can run spatial processes against data on-the-fly from anywhere using standard Web development practices. One place we’ve implemented this is in a prototype developed for the USGS National Hydrography Dataset, allowing editing of data while validating against strict topology. Its current editing tools require up to six hours for a single edit; the effort can be drastically reduced—down to about one minute—using a Web-based editor capable of enforcing a complex set of topology rules across multiple layers.
We are also working to make processing operations easier in browser-based visualizations. Rendering transformations enable just-in-time use of any WPS process as part of a layer’s style. The process is applied on-the-fly to only the area being viewed rather than the full dataset. This capability provides immediate visual feedback without the need to process the entire body of data on a desktop GIS system. NASA’s Global Learning and Observations to Benefit the Environment (GLOBE) Program leverages this feature for fast, dynamic presentation of interpolated surfaces derived from environmental measurements collected at schools around the world.
DM: Is feature parity with ArcGIS Server a long-term goal for the Suite? At version 3.0 is anything “big” missing? Is a feature comparison a meaningful way to evaluate the Suite?
RP: One goal of the OpenGeo Suite is to solve many of the same problems as ArcGIS for Server, though not necessarily by providing the same tools. Both products can publish data from a variety of enterprise databases, including Oracle Spatial and Microsoft SQL Server. Both offer Web services for querying and editing features, publishing and caching map tiles, and running processes on spatial data. Both provide tools for building Web mapping applications or mobile applications.
Their primary differences are in approach. ArcGIS for Server is a server-client system that’s been adapted to accommodate Web services; the OpenGeo Suite was built for the Web from the ground up taking into account some of the best practices in the IT field. Rather than ship a Swiss Army knife with many highly-specialized features, we focus on developing a powerful base of functions into the OpenGeo Suite, then providing tools for building applications to solve specific problems.
In terms of what is missing, we’re actively working on integrating a CS-W catalog service into the OpenGeo Suite and we’re also looking to improve interoperability with desktop-based GIS products such as QGIS and ArcGIS for Desktop.
DM: What single enhancement at release 3.0 is likely to turn the most heads in the geospatial or technology community? Why?
RP: As you can probably tell from my answer above, we’re most excited about bringing processing to the OpenGeo Suite. Having processing in our complete Web mapping solution means customers can work with larger data and perform just-in-time analysis with fewer bottlenecks and errors than with a desktop-based GIS workflow. We believe that exposing data and tools on the Web will provide IT professionals the flexibility to solve unique problems in innovative ways—either by directly invoking existing WPS processes, writing new processes in common languages, using rendering transformations to style data, or in ways we haven’t quite thought of yet.