Ingest large amounts of data from a big data analytics process

MagnetoDB will be useful for interactive access to the results of large analytics processes. For example, suppose a user reduces multiple PBs of raw data into a few 10s of TBs and billions of rows of information. The user may likely want interactive access to that information, to retrieve sets of rows based on different query conditions. In order to support this scenario, MagnetoDB needs to support a process that efficiently imports very large data sets from a big data cluster running map reduce jobs. It should support both full table updates and delta updates to the result data stored in MagnetoDB.