Resilient attachment to provider link state bridging (PLSB) networks

1. A method of interfacing a Client System (CS) with a network domain, the CS including an Aggregator function for load spreading ingress traffic originating from the CS across two or more parallel links of a Link Aggregation Group (LAG), the method comprising:

providing a set of two or more peer edge nodes of the network domain, each one of the set of peer edge nodes including:

a local LAG port hosting a corresponding LAG link connected to the CS; and

at least one port hosting an inter-node trunk connecting the node to each of the other ones of the set of peer edge nodes;

computing a respective path through the network domain from each peer edge node to at least one predetermined destination address, for carrying traffic to or from the CS; and

at each peer edge node, installing forwarding state such that:

ingress traffic received from the CS is forwarded via the path to the at least one predetermined destination address; and

egress traffic destined for the CS is forwarded to the CS via the local LAG port, if the local LAG port is operational, and otherwise via the inter-node trunk to another one of the set of peer edge nodes;

wherein computing a path through the network domain comprises:

computing a set of two or more equal cost paths from any given peer edge node to the at least one predetermined destination address; and

selecting one path from among the computed set of two or more equal cost paths; and

wherein a respective set of two or more equal cost paths are computed for each one of at least two of the set of peer edge nodes, and wherein selecting one of the computed set of two or more equal cost paths comprises:

assigning a respective different rank to each of the involved peer edge nodes, for which two or more equal cost paths were calculated; and

for each involved peer edge node:

assigning a respective different rank to each one of its respective set of two of more equal cost paths; and

selecting the path having a rank matching that of the involved peer edge node.

2. The method of claim 1 , wherein the network domain is a Provider Link State Bridging (PLSB) network domain, and each peer edge node is a Backbone Edge Bridge (BEB) of the PLSB domain.

3. The method of claim 2 , wherein computing a path through the network domain further comprises, at each one of the set of peer edge nodes:

assigning a bearer media-access-control (B-MAC) address to the LAG; and

advertising the LAG in the PLSB domain using the assigned B-MAC.

4. The method of claim 3 , wherein each peer edge node assigns the same B-MAC address to the LAG.

5. The method of claim 4 , wherein the B-MAC address assigned to the LAG corresponds with a MAC Client distributed across the peer edge nodes.

6. The method of claim 4 , wherein advertising the LAG in the PLSB domain comprises:

representing the LAG in the PLSB domain as a phantom node notionally located on the inter-node trunk one hop from each peer edge node, the phantom node having the assigned B-MAC address;

controlling the peer edge nodes to generate Link State Packets on behalf of the phantom node; and

propagating each of the generated Link State Packets through other nodes as if they had been received from the phantom node.

7. The method of claim 3 , wherein the B-MAC address assigned to the LAG corresponds with a MAC address of the local LAG port.

8. The method of claim 7 , further comprising, at each peer edge node:

assigning a respective different Backbone V-LAN ID (B-VID) to the LAG; and

advertising the B-MAC address with the assigned B-VID.

9. A system for interfacing a Client System (CS) with a network domain, the CS including an Aggregator function for load spreading ingress traffic originating from the CS across two or more parallel links of a Link Aggregation Group (LAG), the system comprising:

a set of two or more peer edge nodes of the network domain, each one of the set of peer edge nodes including:

a local LAG port hosting a corresponding LAG link connected to the CS; and

at least one port hosting an inter-node trunk connecting the node to each of the other ones of the set of peer edge nodes;

a respective path through the network domain from each peer edge node to at least one predetermined destination address, for carrying traffic to or from the CS; and

wherein at each peer edge node, forwarding state is installed such that:

ingress traffic received from the CS is forwarded via the path to the at least one predetermined destination address; and

egress traffic destined for the CS is forwarded to the CS via the local LAG port, if the local LAG port is operational, and otherwise via the inter-node trunk to another one of the set of peer edge nodes;

wherein a path through the network domain is computed by a method comprising:

computing a set of two or more equal cost paths from any given peer edge node to the at least one predetermined destination address; and

selecting one path from among the computed set of two or more equal cost paths; and

wherein a respective set of two or more equal cost paths are computed for each one of at least two of the set of peer edge nodes, and wherein selecting one of the computed set of two or more equal cost paths comprises:

assigning a respective different rank to each of the involved peer edge nodes, for which two or more equal cost paths were calculated; and

for each involved peer edge node:

assigning a respective different rank to each one of its respective set of two of more equal cost paths; and

selecting the path having a rank matching that of the involved peer edge node.

10. The system as claimed in claim 9 , wherein the network domain is a Provider Link State Bridging (PLSB) network domain, and each peer edge node is a Backbone Edge Bridge (BEB) of the PLSB domain.

11. The system as claimed in claim 10 , wherein computing a path through the network domain further comprises, at each one of the set of peer edge nodes:

assigning a bearer media-access-control (B-MAC) address to the LAG; and

advertising the LAG in the PLSB domain using the assigned B-MAC.

12. The system as claimed in claim 11 , wherein each peer edge node assigns the same B-MAC address to the LAG.

13. The system as claimed in claim 12 , wherein the B-MAC address assigned to the LAG corresponds with a MAC Client distributed across the peer edge nodes.

14. The system as claimed in claim 12 , wherein advertising the LAG in the PLSB domain comprises:

representing the LAG in the PLSB domain as a phantom node notionally located on the inter-node trunk one hop from each peer edge node, the phantom node having the assigned B-MAC address;

controlling the peer edge nodes to generate Link State Packets on behalf of the phantom node; and

propagating each of the generated Link State Packets through other nodes as if they had been received from the phantom node.

15. The system as claimed in claim 11 , wherein the B-MAC address assigned to the LAG corresponds with a MAC address of the local LAG port.

16. The system as claimed in claim 15 , further comprising, at each peer edge node:

assigning a respective different Backbone V-LAN ID (B-VID) to the LAG; and

advertising the B-MAC address with the assigned B-VID.