Wait, wait... So Apache Spark is basically distributed parallel collections?