Your team is working on a distributed machine learning model that requires iterative processing. Which Spark feature enables the caching of intermediate data in memory to speed up iterative computations?