Why do we need Replication factor at Keyspace level in cassandra, why not at column family level?

Question

Why do we need Replication factor at Keyspace level in cassandra? why not at column family level ?

Answer 1

With the replication factor you specify your data redundancy requirements. The keyspace level is convenient so you don't have to repeat that requirement for every column family you create (and possibly forget). The column families inherit the consistency level of their keyspace. If you have column families with differing redundancy requirements just group them in separate keyspaces. E.g. for your business data you might want a replication factor of 3 whilest for some meta data that you can affort to loose because it rebuilds itself over time, use a separate keyspace with replication factor 1.