AHeise commented on pull request #17345:
URL: https://github.com/apache/flink/pull/17345#issuecomment-958847334


   > @AHeise when you say "module name" are you referring to the Java package 
or the maven artifact? I am happy to go with your advice on the package name, 
we can drop `async` in favour of `unified`. In terms of the module name I am 
also not sure about this:
   > 
   > > flink-connector-aws_2.12
   > 
   > This will result in a bloated uber jar containing all the AWS connectors 
and SDK, and I assume Data/Table API. We need to consider SQL variant too. I 
can see a few options here. Each one is trading off simplicity for user vs jar 
size/dependency chain.
   > 
   > Uber jar:
   > 
   > > flink-connector-aws_2.12
   > > flink-sql-connector-aws_2.12
   > 
   > AWS Service split (we may end up with duplicated AWS SDK if user includes 
multiple `SQL` connectors):
   > 
   > > flink-connector-aws-kinesis_2.12
   > > flink-sql-connector-aws-kinesis_2.12
   > > flink-connector-aws-firehose_2.12
   > > flink-sql-connector-aws-firehose_2.12
   > > etc
   > 
   > API Split:
   > 
   > > flink-connector-aws-table_2.12
   > > flink-sql-connector-aws-table_2.12
   > > flink-connector-aws-data-stream_2.12
   > 
   > AWS Service and API Split:
   > 
   > > flink-connector-aws-kinesis-table_2.12
   > > flink-sql-connector-aws-kinesis-table_2.12
   > > flink-connector-aws-kinesis-data-stream_2.12
   > > flink-connector-aws-firehose-table_2.12
   > > flink-sql-connector-aws-firehose-table_2.12
   > > flink-connector-aws-firehose-data-stream_2.12
   > > etc
   > 
   > I am inclined towards the `AWS Service split` approach. What are your 
thoughts?
   
   I think we need to split by APIs (I don't want to) because `sql` is usually 
a fat jar with shaded dependencies while `datastream` is a normal jar with 
dependency tree.
   
   I'd prefer uber jar iff the dependencies are more or less the same (let's 
say they all use the aws-sdk.jar). If with the new APIs, the dependencies are 
distinct, then we should split them down (e.g. your last option).


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to