System and method for protecting against failure of a TE-LSP tail-end node

1. A method for protecting against failure of a tail-end node of a Traffic Engineering (TE) Label Switched Path (LSP) in a computer network, the method comprising:

learning, at a point of local repair (PLR) along the TE-LSP, reachable address prefixes of next-hop routers from the tail-end node (NNHOPs);

creating a backup tunnel from the PLR to each NNHOP that excludes the tail-end node;

associating each backup tunnel with one or more learned address prefixes; and

in response to failure of the tail-end node, i) performing a penultimate hop popping (PHP) operation on traffic received along the TE-LSP at the PLR, and ii) rerouting the received traffic at the PLR onto an appropriate backup tunnel to an NNHOP corresponding to a destination prefix of the traffic.

2. The method as in claim 1 , further comprising:

computing a shortest path tree (SPT) rooted at the tail-end node to learn the address prefixes.

3. The method as in claim 1 , further comprising:

storing the learned address prefixes in a backup routing table (BRT) at the PLR.

4. The method as in claim 1 , further comprising:

storing the backup tunnels in a backup routing table (BRT) at the PLR.

5. The method as in claim 1 , further comprising:

ceasing to refresh the downstream state of the TE-LSP in response to failure of the tail-end node.

6. The method as in claim 1 , further comprising:

receiving notification that the TE-LSP is to be torn down;

tearing down the TE-LSP; and

tearing down the backup tunnel to each NNHOP in response to the TE-LSP being torn down.

7. A system for protecting against failure of a tail-end node of a Traffic Engineering (TE) Label Switched Path (LSP) in a computer network, the system comprising:

one or more next-hop routers from the tail-end node (NNHOPs); and

a point of local repair (PLR) along the TE-LSP; the PLR adapted to

i) learn reachable address prefixes of the NNHOPs,

ii) create a backup tunnel to each NNHOP that excludes the tail-end node,

iii) associate each backup tunnel with one or more learned address prefixes, and

in response to failure of the tail-end node, iv) perform a penultimate hop popping (PHP) operation on traffic received along the TE-LSP, and v) reroute the received traffic onto an appropriate backup tunnel to an NNHOP corresponding to a destination prefix of the traffic.

8. The system as in claim 7 , further comprising:

a head-end node of the TE-LSP; and

a signaling message sent from the PLR to the head-end node, the signaling message indicating to the head-end node that the PLR is capable of protecting the tail-end node.

9. The system as in claim 7 , further comprising:

a head-end node of the TE-LSP; and

a signaling message sent from the PLR to the head-end node in response to failure of the tail-end node, the signaling message indicating to the head-end node that the PLR is currently protecting the tail-end node.

10. A node for protecting against failure of a tail-end node of a Traffic Engineering (TE) Label Switched Path (LSP) in a computer network, the node comprising:

a network interface to receive traffic along the TE-LSP;

a memory adapted to store i) learned reachable address prefixes of next-hop routers from the tail-end node (NNHOPs), and ii) backup tunnels created to each NNHOP that excludes the tail-end node, wherein each backup tunnel is associated with one or more learned address prefixes; and

a processor adapted to i) perform a penultimate hop popping (PHP) operation on the received traffic, and ii) in response to failure of the tail-end node, reroute the received traffic onto an appropriate backup tunnel to an NNHOP corresponding to a destination prefix of the traffic.

11. The node as in claim 10 , further comprising:

a backup routing table (BRT) within the memory adapted to store learned address prefixes and backup tunnels.

12. The node as in claim 10 , wherein the processor is further adapted to learn the address prefixes reachable from the NNHOPs.

13. The node as in claim 10 , wherein the processor is further adapted to create the backup tunnels to the NNHOPs.

14. The node as in claim 10 , wherein the network interface is further adapted to receive state refresh messages for the TE-LSP from upstream nodes along the TE-LSP, and in response to failure of the tail-end node, the network interface adapted to return the state refresh messages to the upstream nodes without forwarding the state refresh messages downstream to the tail-end node.

15. An apparatus for protecting against failure of a tail-end node of a Traffic Engineering (TE) Label Switched Path (LSP) in a computer network, the apparatus comprising:

means for learning, at a point of local repair (PLR) along the TE-LSP, reachable address prefixes of next-hop routers from the tail-end node (NNHOPs);

means for creating a backup tunnel from the PLR to each NNHOP that excludes the tail-end node;

means for associating each backup tunnel with one or more learned address prefixes; and

in response to failure of the tail-end node, i) means for performing a penultimate hop popping (PHP) operation on traffic received along the TE-LSP at the PLR, and ii) means for rerouting the received traffic at the PLR onto an appropriate backup tunnel to an NNHOP corresponding to a destination prefix of the traffic.