Thanks Xiqian for the proposal, I agree that the multimodal data pipeline is important for us.
+1 for the proposed roadmap and RM candidate. Best, Leonard On Wed, Apr 29, 2026 at 7:05 PM Xiqian YU <[email protected]> wrote: > Hi devs, > > > As Flink CDC 3.6 was released recently, I’d like to propose a discussion > on the next Flink CDC 3.7 release cycle. > > > In this version, we’d like to focus on the following development tasks: > > > 1) Multimodal data support. More data types (OBJECTREF, BLOB, VECTOR, > TENSOR) could be added for unstructured data ingestion sources. Built-in > multimodal processing functions could be added, and sinks like Paimon and > Fluss could write them in a standard way. > > > 2) AI ecosystem integration. Align AI model and function APIs with Flink > SQL edition. A Python SDK is planned for constructing YAML pipelines and > interoperating UDF with the Python ecosystem. > > > 3) Connector ecosystem enhancement. For MySQL CDC, conditional snapshot, > source-side rate-limiting, and binlog deserialization optimization are on > the roadmap. More connectors like HLS (capturing live-streaming media) and > Fluss will be added, too. > > > For developers who are interested in participating and contributing new > features in this release cycle, please feel free to create a task that > targets cdc-3.7.0 to trace your planning features [1]. > > > I'm happy to volunteer as a release manager and open to working together > with someone on this. > > > To ensure that we can complete the above plan, we plan to complete the > development of Flink CDC 3.7 on 2026-06-28, with a 2-month developing cycle. > > > Best Regards, > > Xiqian > > > [1] > https://issues.apache.org/jira/issues/?jql=project%20%3D%20FLINK%20AND%20component%20%3D%20%22Flink%20CDC%22 >
