This specifies whether a specific monitored instance acts as the source of a fault condition being watched for, or as the destination that reacts when the associated source instance reports a fault, the mechanism underlying the unidirectional/bidirectional behavior configured at the group level. An admin setting up a scheme where a downstream port should shut down when an upstream port fails would configure the upstream instance as source and the downstream instance as destination. For example, a downstream access port configured with mode destination would take the group's configured action only when its paired source instance detects a fault, not the other way around.