Ask HN: Are we need a new distributed file system

i was have a project and this projects needs a reliable, fast, distributed file system for storing/accessing big files. (not posix compliant)

So i start to search for already invented wheels, and i found some solutions (HDFS, GlusterFS, ExtremeFS, etc.) but they have a problems like hard to configuration or hard to adding new node or getting slowing when you reach some node number even SPF and/or data loss.

And all file systems needs a effort for customisation using for web (they don't have tunable replication count for file properties and hitting traffic)

finally i convinced my self, we need a new distributed file system which covered all these needs (specially using for web projects, no SPF, minumum configuration, tunable replication, etc.)

but right now i have doubt about it, are we really need a new DFS?