Benchmarking Apache Spark on a Single Node Machine - The Databricks Blog