Break before make forwarding information base (FIB) population for multicast

1. A method of installing forwarding state in a first link state protocol controlled network node having a topology database representing a topology of the network, the method comprising:

detecting whether an “unsafe” condition exists and, when an “unsafe” condition is detected, removing forwarding state for multicast trees originating at a second node in the network;

subsequently advertising a “safe” indication signal to each neighbour node of the first node, the “safe” indication signal comprising a digest of the topology database; and

installing forwarding state associated with a multicast path in a filtering database (FDB) of the first node when a predetermined “go” condition is satisfied, the “go” condition comprising receiving at least one respective “safe” indication signal from at least one neighbour node, each received “safe” indication signal containing a respective topology database digest matching the topology database digest of the “safe” indication signal advertised by the first node.

2. The method of claim 1 , further comprising computing, in response to a change in the topology database, a unicast path from the first node to the second node using the topology database.

3. The method of claim 2 , wherein detecting whether an “unsafe” condition exists comprises detecting a difference in a characteristic of the unicast path to the second node relative to that of a previous unicast path to the second node.

4. The method of claim 3 , wherein the characteristic of the unicast path is a distance to the second node.

5. The method of claim 3 , wherein the characteristic of the unicast path is a route traversed by each of the computed and previous unicast paths to the second node.

6. The method of claim 2 , wherein detecting whether an “unsafe” condition exists comprises detecting whether the computed unicast path would not have been a valid path in the previous topology.

7. The method of claim 1 , wherein the digest of the topology database advertised by the node comprises information representing at least a portion of the topology database for which the unicast path was calculated.

8. The method of claim 1 , wherein the “go” condition comprises receiving a respective “safe” indication signal from each neighbour node of the first node.

9. The method of claim 1 , wherein the “go” condition comprises receiving a “safe” indication signal from each neighbour node of the first node on the multicast path.

10. The method of claim 6 , wherein detecting whether an “unsafe” condition exists comprises, when the computed unicast path would have been a valid path in the previous topology, further determining whether a first neighbour node that is one hop from the first node towards the second node along the computed unicast path in the current topology database is further from the second node than a second neighbour node was in the previous topology database.

11. The method of claim 4 , wherein the “go” condition comprises:

when the distance to the second node is less than a previous distance to the second node, receiving a “safe” indication signal from a neighbour node of the first node that is one hop from the first node towards the second node along the computed multicast path; and

when the distance to the second node is greater than a previous distance to the second node, receiving a respective “safe” indication signal from each neighbour node of the first node that is one hop from the first node away from the second node along the computed multicast path.

12. The method of claim 1 , wherein detecting whether an “unsafe” condition exists comprises receiving a respective “safe” indication signal from at least one neighbour node containing a respective topology database digest that does not match the topology database digest of the “safe” indication signal advertised by the first node.

13. A link state protocol controlled network node, comprising:

a topology database operable to represent a topology of the network;

“unsafe” condition detection functionality operable to detect whether an “unsafe” condition exists;

a filtering database (FDB) operable, when an “unsafe” condition is detected, to remove forwarding state for multicast trees originating at another node in the network; and

communication functionality operable, once the multicast trees originating at the other node are removed to advertise a “safe” indication signal to each neighbour node of the node, the “safe” indication signal comprising a digest of the topology database;

the FDB being operable to install forwarding state associated with a multicast path when a predetermined “go” condition is satisfied, the “go” condition comprising receiving at least one respective “safe” indication signal from at least one neighbour node, each received “safe” indication signal containing a respective topology database digest matching the topology database digest of the advertised “safe” indication signal.

14. The node of claim 13 , wherein the node is operable to compute, in response to a change in the topology database, a unicast path from the node to the other node using the topology database.

15. The node of claim 14 , wherein the “unsafe” condition detection functionality is operable to detect a difference in a characteristic of the unicast path to the other node relative to that of a previous unicast path to the other node.

16. The node of claim 15 , wherein the characteristic of the unicast path is a distance to the other node.

17. The node of claim 15 , wherein the characteristic of the unicast path is a route traversed by each of the computed and previous unicast paths to the other node.

18. The node of claim 14 , wherein the “unsafe” condition detection functionality is operable to detect whether the computed unicast path would not have been a valid path in the previous topology.

19. The node of claim 13 , wherein the digest of the topology database advertised by the node comprises information representing at least a portion of the topology database for which the unicast path was calculated.

20. The node of claim 13 , wherein the “go” condition comprises receiving a respective “safe” indication signal from each neighbour node of the node.

21. The node of claim 13 , wherein the “go” condition comprises receiving a “safe” indication signal from each neighbour node of the node on the multicast path.

22. The node of claim 18 , wherein the “unsafe” condition detection functionality is operable, when the computed unicast path would have been a valid path on the previous topology, to determine whether a first neighbour node that is one hop from the node towards the other node along the computed unicast path in the current topology database is further from the other node than a second neighbour node was in the previous topology database.

23. The node of claim 16 , wherein the “go” condition comprises:

when the distance to the other node is less than a previous distance to the other node, receiving a “safe” indication signal from a neighbour node of the node that is one hop from the node towards the other node along the computed multicast path; and

when the distance to the other node is greater than a previous distance to the other node, receiving a respective “safe” indication signal from each neighbour node of the node that is one hop from the node away from the other node along the computed multicast path.

24. The node of claim 13 , wherein the “unsafe” condition detection functionality is operable to detect reception of a respective “safe” indication signal from at least one neighbour node containing a respective topology database digest that does not match the topology database digest of the “safe” indication signal advertised by the node.