E-MapReduce:Compute

Compute types

EMR Serverless Spark supports the following compute types:

• SQL compute: used for SQL development and queries. If you want to perform a query in the SQL editor, you must select an appropriate SQL compute.

• Notebook compute: used to run notebooks. When you develop a notebook, you must create a notebook compute or select an existing notebook compute based on your business requirements.

• Gateway: used to deploy the Livy service. If you want to submit jobs to EMR Serverless Spark over the Livy interface by using tools such as Airflow and Jupyter, you must specify a gateway compute.

• Spark Thrift Server: a service provided by Apache Spark. Spark Thrift Server allows you to connect to Spark and execute SQL queries based on Java Database Connectivity (JDBC) or Open Database Connectivity (ODBC). This helps you integrate Spark with existing business intelligence (BI) tools, data visualization tools, and other data analysis tools.

• Queue: used to run application jobs and jobs on workflow nodes.

References

• Manage SQL computes

• Manage notebook computes

• Manage gateways

• Manage Spark Thrift Servers

• Manage resource queues