deep storage


Is there documentation describing the fileystem and/or hdfs deep storage. Like, are you writing your our column store. or are you using parquet or orc? I see you have documented an extension for cassandra, but I was sure how you were storing the filesystem or hdfs data.

Thank you!



Here you have documentation about local deep storage and also a part about hdfs deep storage :

Yes, but is there any information about the implementation of the column store. Maybe a white paper?

Oh sorry, I missed the how.
I didn’t read everything but I didn’t see any docs.

But in the github repo, under publications/white paper folders, you have a file named druid.pdf

At the bottom of the 5th page starts a part about storage format

Perfect! Thank you!

The whitepaper is a couple of years old at this point so not everything is still valid.

The most up to date info about segments can be found here: