Essentially by having DynamoDBDataStore implement MapReduceDataStore, data stored within it will inherently be able to be used as a Hadoop InputFormat for distributed processing.