Apache Iceberg REST Catalog: The Complete Guide

The Apache Iceberg REST Catalog is a standardized HTTP API for managing Iceberg table metadata. It is the recommended catalog type by the Iceberg community because it decouples catalog operations from specific implementations — any REST-compliant server (Polaris, Lakekeeper, Tabular, custom) can serve as the backend.

Why REST Catalog?

Catalog Type	Coupling	Multi-Engine	Recommended
REST	Loose (HTTP API)	✅ Best	✅ Yes
Hive	Tight (Thrift)	✅ Good	Legacy
JDBC	Moderate (SQL)	✅ Good	Simple setups
Hadoop/Storage	Tight (filesystem)	⚠️ Limited	Development only

REST catalog advantages:

Interoperability: Any engine (Spark, Flink, Trino, RisingWave) connects via HTTP
Access control: Centralized authorization at the catalog level
Abstraction: Change the backend (Hive → JDBC → custom) without changing clients
Cloud-native: Works naturally with containerized and serverless deployments

REST Catalog Implementations

Implementation	Type	Features
Polaris (Snowflake)	Open source	Full-featured, Snowflake-backed
Lakekeeper	Open source	Rust-based, lightweight
Tabular	Commercial	Managed service
Unity Catalog (Databricks)	Open source	Multi-format support
AWS Glue	Managed	AWS-native
Gravitino	Open source	Multi-catalog federation

Using REST Catalog with RisingWave

CREATE SINK to_iceberg AS SELECT * FROM stream
WITH (
  connector = 'iceberg',
  catalog.type = 'rest',
  catalog.uri = 'http://polaris:8181',
  warehouse.path = 's3://lakehouse/warehouse',
  database.name = 'analytics',
  table.name = 'events'
);

Frequently Asked Questions

Which Iceberg catalog should I use?

REST catalog is recommended for new projects. It provides the most flexibility and interoperability. Use Hive catalog if you have existing Hive Metastore infrastructure. Use JDBC for simple PostgreSQL/MySQL-backed setups.

Does RisingWave support REST catalog?

Yes. RisingWave supports REST, Hive, JDBC, Storage, and AWS S3 Tables catalogs for Iceberg sinks.