This partnership with the Maryland State GIO, the Center for GIS (CGIS) at Towson University (TU) led a project to assess the value of making Maryland’s latest basemap data available in Amazon’s Cloudbased web services. MD iMap is Maryland’s public-facing enterprise GIS infrastructure. We compared the load handling capability of the current MD iMap hosting environment with an Amazon EC2 hosting environment built around a recommended Large server instance. This test intended to quantitatively compare, the Amazon Virtual core EC2 compute units with CPU capacity on MD iMap servers to clarify this important architecting parameter and provide a recommendation for a server instance size suitable as a basis for hosting MD iMap services in EC2. Additionally, we tested the server scaling capabilities for cost of ownership analysis. We determined that a Large EC2 instance was not robust enough to meet the peak demand that MD iMap occasionally responds to. For a variety of methodological reasons we were not able to quantify the differences between current MD iMap server CPU capacity verses EC2 compute units of the Large instance. While an estimated cost of ownership was completed showing MD iMap’s traditionally hosted infrastructure having a lower cost over a 6 year period, there are many ways in which Amazons cost could be brought down by fine tuning testing that was beyond the scope of this project. The most savings in Amazons costs would be wrought by rebuilding GIS services, applications and architecture completely around Amazons hosting services. The project is completed.
Center for GIS at Towson University
Contacts:
Ardys Russakis
410-704-5288
arussakis@towson.edu
Mark Helmken
410-704-5292
mhelmken@towson.edu
Michael Bentivegna
410-704-5299
mbentivegna@towson.edu
Barney Krucoff
410 974-5027
Bkrucoff@gov.state.md.us