Hi Mansi,

Thank you for reaching out and for your interest in the Flink Connector for 
IoTDB 2.X Table Mode project for GSoC 2026. 

It's exciting to see your background in Apache Beam (congrats on those merged 
PRs!) and your familiarity with Java and 
distributed systems—this aligns well with the work involved. It's also great 
that you're already reviewing the existing
flink-iotdb-connector; that's a solid starting point.

To address your questions:

1. We expect the table mode connector to be built as a separate module, similar 
to how the spark-iotdb-connector is structured.

2. For the initial implementation, there's a preference to prioritize the Flink 
Table API. This is the primary integration 
point for many users and will help achieve a minimal viable product (MVP) for 
core read/write operations more efficiently. 
Once that's solid, we can expand to the DataStream API if time allows.

3. This task has had other prospective contributors express interest, but no 
one is currently assigned or actively working on 
it in an official capacity. Per the GSoC process, the scope remains open, and 
there will be a contributor application and 
approval phase later on to select participants.


Regarding a small contribution to get familiar with the IoTDB codebase, that's 
a fantastic idea. You can directly browse 
the open issues at https://github.com/apache/iotdb/issues and pick one to work 
on.
If you have more questions or want feedback on your proposal ideas, don't 
hesitate to follow up. 

Looking forward to seeing your contributions!

Best regards,
Haonan Hou


> On Feb 22, 2026, at 4:27 AM, Mansi Maanu <[email protected]> wrote:
> 
> Hi Haonan,
> 
> I'm Mansi Singh, a Master's student in Information Systems at Northeastern 
> University, and I'm interested in working on the Flink Connector for IoTDB 
> 2.X Table Mode project for GSoC 2026.
> 
> I've been contributing to Apache Beam's Python SDK — my merged PRs include 
> #37672 (fix GroupBy snippet tests) and #37674 (make GCS filesystem lookup 
> lazy to match S3 behavior). I also have experience with Java and distributed 
> systems, and I've been reading through the existing flink-iotdb-connector to 
> understand how the tree mode connector is structured before diving into the 
> table mode work.
> 
> A few questions to help me get started:
> 1. Should the table mode connector be built as a separate module from the 
> existing flink-iotdb-connector, or extend it?
> 2. Is there a preference for targeting Flink's DataStream API, Table API, or 
> both for the initial implementation?
> 3. Is anyone already working on this, or is the scope fully open for GSoC?
> 
> I'd love to make a small contribution to the IoTDB codebase before the 
> proposal deadline to get familiar with the project. Is there a good first 
> issue you'd recommend?
> 
> Thank you for your time!
> 
> Best,
> Mansi Singh
> Northeastern University
> GitHub: https://github.com/MansiSingh17

Reply via email to