Hi, I have a 7206 G1 running 12.3(21) with a PA-A3 interface. The ATM interface has sub-interfaces, aal5snap encapsulated, each one of which corresponds to one PVC and to one IP network (30 or /31 usually), either public or in a VRF. I do not control the other end of the fiber link. The other end of each PVC is usually some device that translates between AAL5SNAP and Ethernet, connected to some customer-owned equipement.
It seems that VLAN tags received on the Ethernet are encapsulated into AAL5SNAP, so the customer should be able to configure 802.1q trunking on the ethernet. Question 1: can I bridge together two PVCs so that the ATM packets that come in on one PVC go right out again on the other without being de-encapsulated, or at least no layer-2 changes? Question 2: can I set up something that permits me to distinguish VLANs on the vlan-into-aal5snap-encapsulated PVCs, seeing each PVC as a 802.1q trunk? I've found LANE, I've found "atm pvc vcd vpi vci", but I don't see how to tie it all together, especially if I want to put each VLAN in a separate VRF. Basically what I want to do is have several PVCs, each containing a 802.1q trunk, and bridge them together, something like this: - vlan 5 on PVCs 50/50, 50/51, 50/52 together into a single layer 2 that has no IP address on my 7206 - vlan 6 on PVCs 50/50, 50/51, 50/52 together into a single layer 2 that has a public IP address on my 7206 - vlan 7 on PVCs 50/50, 50/51, 50/52 together into a single layer 2 that has a private IP address in a VRF on my 7206 so that the network behaves as if the several customer devices on the several PVCs were all connected to a simple switch. Feasible? I'll take partial solutions too . . . -- Thanks Nathan _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
