Skip to main content
When creating internal Iceberg tables, you can use one of RisingWave’s two built-in hosted catalog options. A hosted catalog simplifies setup by managing Iceberg metadata directly within RisingWave, eliminating the need to deploy and manage an external catalog service like AWS Glue or a separate Hive Metastore. This is the recommended approach for getting started with internal Iceberg tables.

​
Hosted catalog options

RisingWave supports two types of hosted catalogs, each based on a standard Iceberg protocol:
  • JDBC Hosted Catalog: Utilizes RisingWave’s internal PostgreSQL-compatible metastore to function as a standard Iceberg JDBC catalog. This is the simplest option with zero external dependencies.
  • REST Hosted Catalog: Provides a REST API endpoint for catalog operations, powered by the integrated Lakekeeper service.

​
Benefits of using a hosted catalog

  • Zero external dependencies: No need to set up and manage AWS Glue, PostgreSQL, or other catalog services.
  • Rapid development: Get started with internal Iceberg tables immediately without complex infrastructure setup.
  • Standard compliance: Both options use standard Iceberg protocols (JDBC and REST) for broad compatibility.
  • External accessibility: Tables created using a hosted catalog can still be accessed by external, Iceberg-compatible tools like Spark and Trino.

​
When to use an external catalog instead

While hosted catalogs are ideal for many use cases, you may want to use an external catalog (like AWS Glue, Hive Metastore, or Nessie) when:
  • You need to share a single catalog across multiple systems (e.g., Spark, Flink, and RisingWave).
  • Your organization has existing investments in or governance requirements for a specific external catalog service.
For details on connecting to your own catalog, see Catalog configuration.
⌘I