Add HDFS StorageBackend implementation

[tigrulya-exe]

This PR adds support for HDFS as a StorageBackend implementation. It also provides Kerberos authentication through the use of a provided keytab and supports asynchronous metric collection based on HDFS client file system statistics.

Users can provide HDFS client configuration in two ways: either by using traditional XML files, specifying their location in the hdfs.core-site.path and hdfs.hdfs-site.path options, or by passing the configuration options as regular Kafka options with the hdfs.conf. prefix.

[jeqo]

Thanks @tigrulya-exe! This looks like a great addition and quite complete coverage of the storage back-end. However, I'm hesitant to move forward on the review as I lack experience on HDFS to be useful on anything apart from the APIs usage.
I'd like to leave this PR open in the meantime to gather feedback and let others to chime in around how to proceed with adding a new back-end.

There are still some work on the project we would like to prioritize before on-boarding a new back-end as well: preparing for Tiered Storage becoming prod-ready in 3.9 or later, and adding docs and release process, etc.

A couple of alternatives while this is open for discussion is to point to your fork (or a separate repo with just HDFS) from our README to let users know there's an HDFS implementation.

Let me know wdyt, and thanks again for your contribution!

[tigrulya-exe]

@jeqo Hi! Thank you for the feedback! I think it's a nice idea to point to our fork with the HDFS storage implementation in your README while this PR is open for discussion :) I don't think we need to create a separate repository just for HDFS, as it could complicate porting future features from the main repository