Update spring-boost or use new framework?

[MisterRaindrop]

Disadvantages:
Spring Boot has heavy dependencies, and many features may not be used. When running independently, it consumes CPU and other resources. The community version has a short maintenance period, requiring frequent upgrades. It heavily relies on the JDK.

Advantages:
Spring has comprehensive functionality.

I have an idea: PXF is just a simple service. Could an extremely minimal framework meet all the operational requirements?

[ostinru]

The community version has a short maintenance period, requiring frequent upgrades. It heavily relies on the JDK.

I don't think that Spring Boot upgrades will cause major issues in the future - they should be relatively easy to update by following upgrade guide.
The biggest problem is Spring 2.x -> 3.x upgrade are not in Spring itself - there will be jump over Java 9 modularization and Java 9 EE to Jakarta 9 EE changes.

It is definitely possible to achieve: https://github.com/arenadata/pxf/pull/118/

Could an extremely minimal framework meet all the operational requirements?

It is good research/PoC project. However I don't see direct impact on users, so it is not on top of my TODO list. In my 'research' pipeline: check Apache Iceberg and IMPORT FOREIGN SCHEMA

[MisterRaindrop]

Integrating a data lake is a good suggestion.
I recommend you pin an issue to outline what we will do next, and then assign tasks accordingly. Like a roadmap？

If assistance is needed, you can @tuhaihe

[ostinru]

and then assign tasks accordingly

Currently I have only Q1 allocated on PXF, so I quite limited in what I will be able to do. Current plan is to collect low-hanging fruits: modernize PXF and verify that there is no regressions.