Today's research increasingly relies on generating, collecting, organizing, and analyzing large amounts of data. Researchers often find themselves needing to work with more data than personal or laboratory computers are equipped to handle, and in some cases, the same scales used by large corporations and government agencies. Scientific Data use cases explain how researchers work with data, including analysis, data management, and data sharing.
These use cases describe how scientists analyze large amounts of scientific data. The data is typically large in volume (larger than one would use on a personal or business computer), but is organized and stored in different ways for different kinds of research. Data might be generated by a single source (a large simulation, for example) or it might come from many sources (observational results from many different instruments or different research teams). The methods of analysis also vary from field to field and problem to problem.
These use cases describe how researchers manage collections of data for shared use or for their own re-use over time. The use cases range from a single research project managing and organizing its own data, to several related projects using each other's data, or to data being prepared for future use in applications that haven't been imagined yet.
|Use Case ID||Title||Use Case Description|
|DM-01||Share a common repository of data with a distributed user community|
|DM-02||Data Management for Distributed Simulation and Analysis|
|DM-03||Shared use of large-scale/streaming sensor input data|
|DM-04||Migration of data associated with change of primary computational site|
|DM-05||Manually create metadata for a data object|
|DM-06||Run a researcher-supplied tool to generate metadata for data objects|
|DM-07||Automatically extract metadata from data objects|
|DM-08||Store metadata for later use|
|DM-09||Search metadata for specific objects of interest|
|DM-10||Add metadata search features to an application|
|DM-11||Post-allocation data access on XSEDE SP resources|
These use cases describe the most common scientific data visualization methods. This is an evolving field of work, since the ability of desktop computers to visualize data is improving rapidly. Nevertheless, it is still not unusual for researchers to need to visualize data at scales that exceed their local resources, requiring them to use high-performance and high-throughput computing resources to advance their work.