Ask HN: Are we need a new distributed file system
1 comments
There are definite holes in what can be done right now. Some projects give you all of the raw tools but none of the
automatic management you need to keep something running. Others scale poorly in terms of human costs when it comes time to add new machines.
If you agree, then throw money at me and I'll build one for you. No, I'm not joking. Contact info is in my profile.
If you agree, then throw money at me and I'll build one for you. No, I'm not joking. Contact info is in my profile.
yes, i agree and if had $1000000, i should spend for that but i haven't.
That's a good sum. You don't have to pay me all at once.
if someone do that properly i guess they will be rich.
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?