Geographic Information Systems at University of Delaware
Traditional geospatial workstations
The Library SMDC lab is equipped with ArcGIS software, suitable for most GIS tasks
The Research Computing Lab in Smith Hall is equipped with all GIS packages licensed by UD (ArcGIS, IMAGINE, ENVI+IDL, FME) as well as the most popular open source GIS packages.
High performance processing of geospatial data
Amazon EC2 is recommended for one-off, memory, and compute intensive tasks that have not been optimized for parallel or distributed computing. IT CS&S can assist with deploying these systems, which are conveniently and rapidly provisioned/released for one-off jobs. The Esri software on the machine (e.g., ArcGIS Server or SDE) is available as part of our site license and therefore often at no extra cost, but Amazon will charge based on use. More information available on request
The Mills Community cluster provides exciting capabilities for ongoing or very large, compute intensive jobs. Jobs must be optimized for parallel or distributed computing to achieve much improvement over a traditional system. We are looking into optimizing ArcGIS and IDL optimizations for this platform, primarily through Python. Matlab is the most mature analytical computing package on the Mills system, and can be used for some types of geospatial analysis.
Serving geospatial data, databases, services, and applications
ArcGIS Online is an easy, intuitive, free platform for sharing many geospatial data types. See below for instances where cloud-based ArcGIS server is a better option.
Cloud-based ArcGIS Server is a new service we’re providing at UD. This service would replace a physical ArcGIS Server or SDE instance, and can provide substantial resource savings and ease of use. This option is useful for large datasets, publishing geoprocessing services, spatial database needs, or where there are intellectual property or access considerations. It is also a much less expensive option than Amazon EC2 over the long-term.
Google Fusion is particularly adept at generating choropleth maps from tabular data based on a boundary-type file
Google My Maps is a consumer-focused option which is the least capable/private of all these options, but is nevertheless preferred in some cases where ease-of-use and interoperability trump other considerations.
UD-IT CS&S • University of Delaware • Newark, DE 19716