Storing data in multiple shards is same as replication factor. We can address this concern by using right term as "Replication factor" !
Not sure they are replicating here. This is just distributing data. Replication is mentioned in the dynamo paper and is a feature of HDFS.