Technique for protecting communication traffic in a connection having redundancy

1. A method for protecting traffic in a Layer 2 (L2) communication network between a first and a second local nodes (CE 1 , CE 2 ) interconnected via a dual homed structure, wherein said structure comprises: four provider edge nodes PEs, four communication lines and four local connections, a first pair of the provider edge nodes (PEs A,C) being respectively connected by a first couple of local connections from among the four local connections (CE 1 -PEs A,C) to the first local node (CE 1 ), a second pair of the provider edge nodes (PEs B,D) respectively connected by a second couple of local connections from among the four local connections (CE 2 -PEs B,D) to the second local node (CE 2 ), two distant communication lines (P-PW) from among the four communication lines each interconnecting two provider edge nodes PEs belonging to different pairs, and two local communication lines (S-PW) from among the four communication lines each interconnecting two provider edge nodes PEs belonging to the same pair;

the method comprising the steps of:

a) in a fault-free state, carrying traffic from the first local node (CE 1 ) to the second local node (CE 2 ) by dividing it into two traffic flows at the first local node (CE 1 ) and respectively carrying the two flows via the first couple of local connections, via the provider edge nodes PEs of said first pair, via said two distant communication lines, via the provider edge nodes (PEs) of said second pair, and via the second couple of local connections, while carrying said two flows via both the provider edge nodes (PEs) of said second pair;

b) in a faulty state, when a failure occurs in at least one element of a group comprising a local connection (CE 1 -PE) and a provider edge node (PE):

propagating Layer 2 fault indications (L2 OAM tools) in the dual homed structure and processing thereof at the PEs so that as a result of said failure, the failed element, the remaining element of the group associated with it and also the local and the distant communication lines associated with said PE element of the group become blocked for said two traffic flows in the dual homed structure, while a specific one of said first and second local nodes CE 1 and CE 2 , being remote from the failure, remains unaware of the failure;

performing switching at the local node being closer to the failure, so that both said two flows are directed to unblocked members of the dual homed structure, and

rearranging said traffic flows based on the physical and said Layer 2 fault indications, so as to bypass all said blocked members of the dual homed structure by utilizing the remaining ones to carry both of said two flows,

while both of the flows, arriving from the first local node (CE 1 ) to the second local node (CE 2 ) upon the mentioned failure, pass through both provider edge node (PEs) associated with the second local node (CE 2 ) in case they are both functional, thereby allowing the PEs associated with the second local node to learn MAC addresses sent by CE 1 .

2. The method according to claim 1 , comprising carrying the two traffic flows in the fault-free state in mutually opposite directions via the local communication line (S-PW) between the nodes (PEs) of said second pair, so as said PEs of the second pair both learn the MAC addresses sent by the CE 1 .

3. The method according to claim 1 , wherein the communication network is VPLS or Ethernet-VLAN.

4. The method according to claim 1 , wherein said two distant communication lines and said two local communication lines are pseudo wires (PWs) or Ethernet-VLANs.

5. The method according to claim 1 , comprising performing load balancing at the CE 1 s when carrying the two traffic flows via the first couple of local connections.

6. The method according to claim 1 , where the L2 fault indications in the form of OAM tools reflect failures of the dual homing structure and are processed at L2 by said PEs.

7. The method according to claim 1 , wherein MAC address learning is performed as follows: in the fault free state, a provider edge node PE receiving a new MAC address from a local communication line (S-PW) learns said address and associates the address with its suitable distant communication line (P-PW), wherein the association is performed: either by listing the address as if it was learned from said distant communication line (P-PW), or by listing the address as if it was learned from the local communication line (S-PW), but actually forwarding traffic as if packet was learned from the P-PW, while the PE flushes MAC addresses whenever a fault indication (FI) received over the S-PW changes from fault-state to fault free and vice versa.

8. The method according to claim 7 , wherein in case said provider edge node (PE) receives fault indications (FI) over its local communication line (S-PW), it starts associating new MAC addresses arriving from said S-PW with its close local node (CE).

9. A method for protecting traffic in a Layer 2 (L2) communication network between a first and a second local nodes (CE 1 , CE 2 ) interconnected via a dual homed structure, wherein said structure comprises: four provider edge nodes PEs, four communication lines and four local connections, a first pair of the provider edge nodes (PEs A,C) being respectively connected by a first couple of local connections from among the four local connections (CE 1 -PEs A,C) to the first local node (CE 1 ), a second pair of the provider edge nodes (PEs B,D) respectively connected by a second couple of local connections from among the four local connections (CE 2 -PEs B,D) to the second local node (CE 2 ), two distant communication lines (P-PW) from among the four communication lines each interconnecting two provider edge nodes (PEs) belonging to different pairs, and two local communication lines (S-PW) from among the four communication lines each interconnecting two provider edge nodes PEs belonging to the same pair,

wherein, in a fault-free state, the method comprises carrying traffic from the first local node (CE 1 ) to the second local node (CE 2 ) by dividing it into two traffic flows at the first local node (CE 1 ) and respectively carrying the two flows via the first couple of local connections, via the provider edge nodes (PEs) of said first pair, via said two distant communication lines, via the provider edge nodes (PEs) of said second pair, and via the second couple of local connections, while carrying two said flows via both the provider edge nodes (PEs) of said second pair, and

wherein in case of failure in one of the distant communication lines, said distant communication line becomes blocked for said two traffic flows, and the method comprises:

rearranging said two traffic flows based on the physical and said Layer 2 fault indications so as to bypass said blocked distant communication line by utilizing the remaining one to carry both of said flows, and

passing both of the flows arriving from the first local node (CE 1 ) to the second local node upon the mentioned failure, through both provider edge node (PEs) associated with the second local node (CE 2 ), in case both of said PEs are functional, thereby allowing said PEs to learn shared MAC addresses.