About Swift, I thinks is very specific for openstack, since I am using a different cloud (private cloud).
How much control do you have over this private cloud - what kind of technology are we talking about here. Because e.g. OpenStack can of course also be run on-premise/in a private cloud.
The challenge I am working is to write blocks, send it to some cloud and delete from my filesystem. When I want to read the files I will restore these blocks from the cloud.
If I understand you correctly, you're looking not so much to write a client to connect to a blobstore, but to write your own blobstore implementation?
jclouds is mainly a *client* to talk to blobstore implementations. The filesystem provider is indeed an exception here, but (as far as I understand, at least) it's intended to be a useful test provider you can use to validate an application that talks to a blobstore using jclouds.
If you're looking for a production-grade blobstore implementation, the links Andrew G mentions are worth looking at. I would echo his comment and strongly recommend that you look for an existing blobstore implementation (such as OpenStack) that meets your requirements, rather than trying to write your own.
Regards ap
