Couch DB can be used as a state database for Hyperledger fabric. As per documentation and from my experience, I understand that couchDB contains only the World state data and not the blockchain data. Also we could use things like design documents, views, pagination, indexes etc. Also couchdb needs to run in a separate container and we map the couchDB container port to the peer.
My question is that if we have terabytes of data in the world state, how does this affect scaling up in couch DB? Can a single container handle this data without taking a performance hit? It is written in the couch DB documentation that when the system grows large enough, a sensible thing is to add more nodes to the database. How do we achieve this in hyperledger fabric?
