We have built a decent SMACK stack (currently without Akka, but that is another question) for a data ingestion system with querying capability.
Now if we want to expose this platform to different users--- say different verticals or applications --- that would want to restrict access between data being ingested, is there any best-practices out there.
I can see some ACL support for Cassandra and even Kafka, but was thinking of completely separating and bringing up a different stack (using mesos/DC-OS) for each application. There is a bit of control overhead, but if data is completely isolated between apps then this is fine.
A close second thought is what to do when the data is NOT completely isolated and you want to restrict access based on credentials. That would require some thought --- any experiences out there?
thanks.
