Using the history of ARM data files, we designed and demonstrated a data versioning paradigm that is feasible. Assigning versions to sets of files that are modified with some special assumptions and domain specific rules was effective in the case of ARM data, which has more than 5000 datastreams and 500TB of data.
Revised: August 19, 2015 |
Published: June 29, 2014
Citation
Macduff M.C., B. Lee, and S.J. Beus. 2014.Versioning Complex Data. In IEEE International Congress on Big Data (BigData Congress 2014), June 27-July 2, 2014, Anchorage, Alaska, 788-791. Piscataway, New Jersey:Institute of Electrical and Electronics Engineers.PNNL-SA-99188.doi:10.1109/BigData.Congress.2014.124