A media company runs an existing Apache Spark and Hadoop cluster on-premises to process large batches of log data. They want to move these open-source workloads to Google Cloud with minimal changes to their existing Spark jobs, while offloading cluster management. Which Google Cloud service should they use?