Engine model
The managed engine includes the services required to accept SQL, execute streaming computations, coordinate metadata, and compact persistent state. StreamNative configures, deploys, and monitors these services as one SQLWorkspace engine. RisingWave uses two managed persistence resources:Metadata Store
Stores engine metadata in Amazon RDS for PostgreSQL.
Object Storage
Stores durable streaming state in Amazon S3.
During Private Preview, managed metadata and object storage are supported only for SQLWorkspace placements on AWS. Other cloud providers are not currently supported.
Query interfaces
You can submit SQL to the same engine through either interface:
PostgreSQL wire compatibility does not mean that every PostgreSQL server feature is available. RisingWave syntax, system catalogs, extensions, and SQL behavior can differ from PostgreSQL.
Work with managed topic sources
StreamNative periodically discovers eligible topics from a SQLCatalog and creates missing source objects in its engine database. Always reference topic-derived schema and source names with double-quoted identifiers. This example uses a Pulsar topic inpublic/default:
RisingWave documentation
- Source, Table, Materialized View, and Sink explains the roles of RisingWave’s core SQL objects.
- Data processing in RisingWave introduces ad hoc queries and continuously maintained streaming computations.
- CREATE MATERIALIZED VIEW provides the complete materialized-view syntax and behavior.
- Joins covers RisingWave join patterns.