Data frames do work well. You are correct that it is mainly an abstraction on top of the existing spark RDD representation. The language is Scala, though a slig

Data frames do work well. You are correct that it is mainly an abstraction on top of the existing spark RDD representation. The language is Scala, though a slightly strange string/macro-based sql-ish API. Perhaps you can tell from that description I'm a bit luke-warm on it. There are a few big benefits to immutability. Redundancy and resilience is a big one, but you also get some improved performance because the compiler and planner can more aggressively optimize, and do things like predicate push-down since it doesn't have to materialize everything. It also gets you a lot of the more general programming benefits of immutability/FP in preventing bugs and such. I somewhat agree with your point about data volume. The situations that require Spark are relatively infrequent. But they are becoming more common. A lot of data is required when you start getting into building more complex machine learning models. But for a lot of your run-of-the-mill regression or basic understanding analysis, Spark is way overkill.