RSVP/Backup Tunnels

Figure 1 illustrates the RSVP/Backup Tunnel restoration scheme. This restoration scheme is based on the MPLS protection scheme introduced in [6]. In this approach, the failed headend⁶ informs the ingress node (which is $f$ hops away from it) about the failure through a FIS message. The ingress node sets up the alternate path by sending a request to the first node on it. Each node forwards the request downstream without actually waiting to complete the cross connect. The egress node returns an acknowledgement which travels the same route upstream. Each node forwards it (upstream) as is, with a positive acknowledgement in it if the cross connect at its site was successfully completed. If not, it puts a negative acknowledgement. When the acknowledgement reaches the ingress, it knows if the path is setup by examining the type (positive or negative) of the acknowledgement. The time required to complete each phase of the restoration process is as follows:

$$t_{detect} = t_{LOL}$$ (4)

$$t_{FIS} = f \cdot d_{DCN}(h)$$ (5)

$$t_{setup} = 2 \cdot l_a \cdot d_{DCN}(h)$$ (6)

$$t_{restore} = t_{LOL} + f \cdot d_{DCN}(h) + 2 \cdot l_a \cdot d_{DCN}(h) + t_{switch}$$ (7)