C2D output in bucket

[bogdanfazakas]

configure free/paid compute with a param on startCompute that accepts a bucket id to store the results
(store files not output archive)
make it usable for a next c2d job input / dataset