A company uses Apache Hadoop and Spark on-prem. The infrastructure is complex and not scalable. They want to reduce operational complexity but keep data processing on-premises. Which solution will meet these requirements?