Setting up ad hoc (aka standalone) Spark cluster with a bunch of machines you have control over is ridiculously trivial task though. It's as easy as running spa

Setting up ad hoc (aka standalone) Spark cluster with a bunch of machines you have control over is ridiculously trivial task though. It's as easy as running spark --master=x where you designate one machine as master. All others started with --master=x become slaves of x. Then you just submit jobs to x and that's all.