Date of Award
Master of Arts
Dr. Kathleen Baker
Dr. Chansheng He
Dr. Lei Meng
Masters Thesis-Open Access
Three separate storage technologies able to serve gridded data were selected for comparison of performance in terms of providing speed and expandability to a crop disease forecasting system. The three storage technologies chosen were PostgreSQL (a relational database management system), MongoDB (NoSQL system), and netCDF files. Speed tests were performed for each by running two different crop disease risk forecasting models requiring data of different spatiotemporal resolutions. Multiple trials were done using different storage hardware. Systems were then qualitatively compared for expandability by noting the process involved in adding successive crop disease forecasting models.
It was found that due to different respective limiting properties of each implementation of all three storage technologies the speed differences using traditional storage hardware were few. Given this, it would be possible to further finetune a system using netCDF files for speed gains. Qualitative notions of expandability featured by the different storage technologies then become a significant factor when making a choice between the three to use for a crop disease forecasting system. Both PostgreSQL and MongoDB storage technologies offered better expandability in terms of difficulty of adding additional models compared to the system using netCDF files.
Roehsner, Paul J., "Data Storage Alternatives for a Gridded Crop Disease Risk Forecasting System" (2013). Master's Theses. 157.