A failure along the path of a signaled RSVP-TE LSP can remain undetected for two minutes or longer (reservation or RESV time-out). During this time, bandwidth is held by the non-functioning LSP on the nodes downstream from the point of failure along the path with the intact state. If this bandwidth is needed by head-end tunnels to signal or re-signal LSPs, tunnels may fail to come up for several minutes, thereby negatively affecting convergence time.

Hello messages enable RSVP nodes to detect when a neighboring node is not reachable. When the RSVP-TE Hello protocol notices that a neighbor is not responding, it treats it as a neighbor down case (link layer communication failure) and deletes the LSP state or reroutes it, based on the type of LSP. This action frees the node's resources so they can be reused by other LSPs.

A Hello message is sent out periodically to each RSVP peer on a link. If no response is received from the peer within a specified period, then the peer is announced "dead" (down). RSVP LSPs going over that peer must either be torn down or re-routed, based on the nature of the LSPs.

This Hello mechanism is intended for use between immediate neighbors. Hello processing between two neighbors supports independent selection of configurations of failure detections intervals.

The configuration of Hello message is completely optional. All the messages may be ignored by nodes which do not wish to participate in Hello message processing.

By default, this feature is disabled.