Architecture¶

Denodo Embedded MPP architecture¶

Denodo Embedded MPP is shipped with an Embedded Hive Metastore that acts as a repository of metadata mapping Parquet files, Iceberg tables or Delta Lake tables - from S3, ADLS, GCS or HDFS - to Hive, Iceberg or Delta Lake tables.

This Embedded Hive Metastore stores the metadata in an Embedded PostgreSQL. Users can also choose to use an alternative external database, (PostgreSQL, MySQL, SQL Server or Oracle), to work with the Embedded Hive Metastore.

When using the Embedded PostgreSQL, the cluster uses a Kubernetes Persistent Volume, to ensure the persistence of the metadata. This way, all the cluster services: Presto, Hive Metastore and PostgreSQL, can be temporarily down and we will not lose any of the table definitions that the Denodo Embedded MPP relies upon.