then other nodes that depend on communication with the afflicted node to execute a transaction need not wait for the failed node to recover back to its original state-rather they can make requests to the replica node for any data needed for the current or future transactions.