I have an API developed with Java and Jersey which operating high amounts of files and data-base related data. Currently I am storing files in S3 on Amazon. Previously, I just stored files in the same file system where tomcat was located so it was hard to achieve replication and well-structured design of system.
My aims:
- Fast I/O (let's imagine bandwidth will not affect this)
- Replication
- Scalability
- It should beLinux-based solution
I need the solution separated from business logic in network. It can be other related services will use it.
Could you please suggest how to achieve that points? Do someone had such experience? For now, I took a look to Samba on Linux. But not sure I am on the right way.
Should I clarify something?