This component is used for snapshotting of the in-memory state of a Hadoop cluster and to keep track of write-ahead logs and transactional attributes. It is often used to replicate the cluster state. This is not a backup node for the NameNode server, but is for snapshotting purposes.