It uses Simple Single Datacenter Clusters. It places the first Replica on a node determined by the Partitioner. Additional Replicas are placed on the next nodes in clockwise (in a Ring) manner without considering Rack or Datacenter location.
