Spark built in connector for Druid

[averma111]

Motivation

In my current project there is a requirement to load HDFS data combined with Db2 data which is huge in number. The data volume is almost 1 TB per day. We need to join these two huge data sets and the perform business logic on top if to create daily report for insights.

Proposed changes

Today we need to use Tranquilizer Beam, storm or any other middle ware with spark to push the data into Druid. If we have spark connector directly connection we can all the transformation directly into spark push it like we have kafka indexing services.

I am not very professional in Java tech stack but learning , hopefully if no is working on this proposal one day I will create it.:)

Thanks,
Ashish

[mangrrua]

Check issue #9780 for batch read/write operation