GeoWorld

GeoWorld September 2012

Issue link: https://read.dmtmag.com/i/84685

Contents of this Issue

Navigation

Page 19 of 32

However, the more VTA developed under this approach, the more it realized the limitations, such as lack of control on the basemap, impaired performance for large overlays, and not being able to run custom- designed SQL queries against Google's database or VTA's overlays. Most importantly, VTA realized that for Web-GIS applications to be the interface for an enterprise GIS it envisioned, the Google API- and JavaScript-based approach is too lightweight to handle the task. Instead, a multitier infrastructure including a database server, Web server and Web client is required. Besides such limitations, VTA's transition was moti- vated by the API's success—introducing the Google Map API started a frenzy of GIS mashup applica- tion development in the online community. The key to the API's success is the same for two famous crowdsourcing OpenStreetMap. Internet companies: Wikipedia and The philosophy at work is "empowering"—the type that "empowers others without losing its own power." By providing information just a click away, companies can tap into the momentum of public enthusiasm and creativity. In turn, this enables companies to provide better information and tools to attract more users. It works like a perpetual engine, with a self-sustainable mechanism built in. VTA saw an opportunity to collaborate with the online development community. After all, VTA creates and maintains a large amount of data and frequently gets e-mails from enthusiastic individual developers requesting data to develop online planning or mobile applications. Providing VTA's GIS data through Web- service layers (an API is basically a Web service) enables the sharing of GIS resources with the online development community and the world in general without losing control of those resources. To generate Web-service layers from GIS data, a map server was added to the multitier infrastructure. A Database Backend Like the foundation of a building, a robust database backend is essential to a Web GIS. Several DBMSs were weighed for suitability, and the final choice was between the proprietary Oracle system or an open- source PostgreSQL system. Oracle was the existing system, but that didn't mean there would be no additional costs. It was purchased several years ago with about 50GB of existing data. Loading GIS data into the system could easily double or triple the size. To handle data of a much larger size and more frequent and complicated queries, the exist- ing system needed to be upgraded. The additional cost could be prohibitive. In the long term, open-source PostgreSQL seemed a more cost-efficient solution. Although some in the agency had concerns about PostgreSQL's open-source nature, its advocates argued that being open source doesn't mean being sloppily programmed or loosely maintained; "open" is more in the sense that the software's source code is open for modification, and this doesn't necessarily open Pandora's Box. Based on documents on PostgreSQL, the software enforces as strict a version control as its proprietary counterparts. The GIT (a free and open-source, distrib- uted version-control system) is used to track a com- plete history of revision. With strict version control, the open nature opens opportunities for improvement without jeopardizing integrity. In essence, the nature of open source is PostgreSQL's advantage. Passing the Test To further dispense any skeptical feelings about PostgreSQL, a virtual server was set up to be a test environment. The first project employing PostgreSQL processed datasets from Census 2010 and American Community Survey (ACS). In these datasets, the geospatial data for different geographical entities, such as Census Tracts, Block Groups or Blocks, come as separate Shapefiles. Unlike geospatial data, the attribute data in ASCII text format are split by data segments, each with all the geographical entities in the state of California bundled together. Figure 1. Within an ACS-based Web-Map Information System, users can pan, zoom, query and turn layers on or off. SEPTEMBER 2O12 / WWW . GEOPLA CE . COM 19

Articles in this issue

Archives of this issue

view archives of GeoWorld - GeoWorld September 2012