At ebay classifieds we started running our infrastructure on top of OpenStack.
The openstack project comes with its own amazon-s3-like filesystem, known as Swift. It's built for scale and optimized for durability, availability, and concurrency across the entire data set. Swift is ideal for storing unstructured data that can grow without bound.
We would really like to be able to use it within flink without Hadoop dependencies, as a sink or for storing savepoints etc
I've prepared a pull request that adds support for it. It wraps the hadoop support for swift in a way that is very similar to the way the s3 connector works.
You can find out more about the underlying hadoop implementation at https://hadoop.apache.org/docs/stable/hadoop-openstack/index.html
Pull request : https://github.com/apache/flink/pull/5296