In a CDP public cloud deployment, DAS is available as one of the many Cloudera Runtime services within the Data Engineering template. To use DAS, create a Data Hub cluster from the Management Console, and select the Data Engineering template from the Cluster Definition dropdown menu.
Each Data Hub cluster that is deployed using the Data Engineering template has an instance of Hive Metastore (HMS), HiveServer2 (HS2), and DAS. The HMS communicates with the Hive RDS in the Data Lake for storing the metadata information. It also communicates with Ranger in the Data Lake for authorization. Hive is configured to use the HMS in the Data Hub cluster, and Ranger in the Data Lake cluster.
DAS comprise of an Event Processor and a Webapp. The Event Processor processes Hive and Tez events, and also replicates the metadata. The Webapp serves as the DAS UI. It serves data from the local Postgres instance which is populated by the event processor, and uses the HiveServer2 to run Hive queries triggered through the Query Compose page. The Event Processor communicates with the Data Hub HS2 instance for the repl dump operation, and reads the events published by Hive and Tez from the cloud storage (S3 or ADLS).
The following diagram shows the components inside the Data Hub cluster which use the shared resources within the Data Lake, such as Ranger, Hive RDS, and the cloud storage: