Which method is used to implement Spark jobs?
in memory of all workers
in memory of the master node
on disk of the master node
on disk of all workers