GitHub user TKilome created a discussion: [DISCUSS] Is a Paimon REST Catalog
Server still necessary when Gravitino is used as the unified access layer?
I am evaluating the recommended architecture for integrating Apache Paimon
with Gravitino.
If users access Paimon through Gravitino connectors, the request path is:
```text
Flink / Spark
-> Gravitino connector
-> Gravitino Server
-> Gravitino Paimon catalog
-> Paimon backend
```
In this model, Gravitino can provide a unified metadata entry point,
authentication, authorization, audit, and catalog governance.
My question is: in this architecture, is it still necessary to implement or
deploy a separate Paimon REST Catalog Server?
More specifically:
1. If Gravitino already exposes Paimon metadata through its own REST API and
engine connectors, what additional value does a Paimon REST Catalog Server
provide?
2. Should the recommended deployment be:
Flink / Spark -> Gravitino -> Paimon filesystem / Hive / JDBC backend
instead of:
Flink / Spark -> Paimon REST Catalog Server
3. Is the Paimon REST Catalog Server mainly useful for clients that do not
integrate with Gravitino directly, or for compatibility with the Paimon REST
catalog protocol?
4. If a Paimon REST Catalog Server is still useful behind Gravitino, should
it be treated only as an internal backend service, with users accessing Paimon
only through Gravitino?
5. From the Gravitino community's perspective, is there any plan or
recommendation around implementing a Gravitino-backed Paimon REST Catalog
Server?
Thanks.
GitHub link: https://github.com/apache/gravitino/discussions/12161
----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]
