Regarding how refer to Sunile. Pig is nice and flexible, Hive is good if you know SQL and your RFID data is already basically in a flat table format, Spark also works well ... But the question is if you really want to process 100GB of data on the sandbox. The memory settings are tiny there is a single drive data is not replicated ... If you do it like this you can just use python on a local machine. If you want a decent environment you might want to set up 3-4 nodes on a VMware server perhaps 32GB of RAM for each? That would give you a nice little environment and you could actually do some fast processing. ... View more