Font Size: a A A

A distributed approach to managing large simulation data sets

Posted on:2006-08-20Degree:M.SType:Thesis
University:Michigan State UniversityCandidate:Connelly, Brian DFull Text:PDF
GTID:2458390008956677Subject:Computer Science
Abstract/Summary:
SimDB is a project to develop a distributed system that stores and analyzes large simulation data sets from molecular dynamics simulations. Using the resources of all the machines on the network, it is able to store a large amount of simulation data (trajectories), so that interested groups may work with these data. Because it would take a long time to transmit these data over the Internet, users instead specify how they would like to analyze the data using a rich set of analysis functions. The analysis is then done online, and the resulting data are sent to the user. This data requires considerably less storage space than the raw data, so transferring it requires less time. SimDB also stores these analysis results, thereby offering researchers the further benefit of accessibility: Stored analysis results may be used to satisfy future queries without the necessity of re-analyzing the data.
Keywords/Search Tags:Large simulation data sets, Analysis results
Related items