Failover is the implementation of mirrored servers that automatically switch to a clone upon the failure or abnormal termination of the Engine. After being configured, failover happens without human intervention. When the clustering ring is in failover mode, the engine doing the processing operates the same as if it were a standalone, with the added functionality of keeping its place in the ring. If the first engine fails, this is noticed within the ring and another engine will initiate the restart process, pull all the work from the state database, and continue with the processing.