We are working on both replication (for high availability and we may use the Raft consensus) and distribution (sharding to scale horizontally and keeping low latency)