How does AWS Glue crawler work and is like a Hive or Spark tech - any open source project is used underneath. I am looking to build something similar on premise