CDF and Granite : A Comparison Master ' s Project Report

CDF and Granite: A Comparison Master's Project Report Lorna Ellis May 2004 Abstract Because large scientific datasets are often stored in files, efficient data analysis depends on efficient I/O and data manipulation tools. The Granite Scientific Database System is a system that incorporates database methodology and spatial semantics to support retrieval and visualization of scientific data. CDF (Common Data Format) is a data format used by many scientists to store their data. CDF has its own access and manipulation routines. This paper discusses the two systems and compares the speed with which they are able to retrieve data from files. Introduction With the continuing increases in the size of many scientific datasets, systems that provide efficient and flexible means of accessing and manipulating data become more and more important. Granite is a relatively new system that combines a database model with spatial semantics to provide multiple resolutions of scientific data. CDF is a wellestablished data format used by many scientists to store their data. Comparing the two offers insights into the strengths and weaknesses of both, and offers scientists and others who analyze scientific data a framework for deciding which system might work best for them. It is also possible that insights gained from this comparison could lead to conceptual changes in future releases of CDF. Granite: Granite is a data model and system that focuses on “retrieval and visualization of gridded scientific data at multiple resolutions” [1]. The prototype system