Enable data generation on GCS, S3 and Azure

[sameerz]

nds_gen_data.py enables data generation locally and on HDFS. Allow options to generate data on GCS, S3 and Azure (in that priority order).

[wjxiz1992]

I think about the Spark way to generate data a while and I don't have a clear picture how this should be done, but this pattern seems very likely to be a SparkStreaming scenario: TPC-DS tool keeps generating data(data source) , Spark should receive them(receiver), it will last for a while (streaming).