Hello,
I am currently deploying an infrastructure based on Q-in-Q tunneling.
On one side, I have my core network : a VSS of 4500-X Catalyst which act as my L3 point (it will be doing all the routing once the network is fully deployed).
On the other side, I have several remote sites connected to a L3 WAN with a old Nortel switch.
I have a brand new LAN-to-LAN link (L2) setup by my ISP which I want to use, connecting my core to each of my remote sites (much like a hub&spoke topology), essentially doing away with my actual L3 WAN. To do this I plan to use Q-in-Q tunneling to transport my vlans from the remote site to the core site. The problem is, my Nortel switch does not handle vlans, nor trunking, making it difficult to use the LAN-to-LAN link with Q-in-Q.
Therefore, I will have to send all of my remote site's untagged traffic down this link, essentially tagging all of this as one Q-in-Q vlan (in fact, tagging the native vlan).
However, when I will decapsulate my frame on the core site, I will have all this untagged traffic and no way to dispose of it. Once the traffic reaches my core site, I am supposed to route it somewhere (it doesn't matter where), but I can't do so without knowing the kind of traffic it is.
Has anyone encountered this particular problem? I can't be the only one having this kind of question because this is a rather simple architecture, so I am curious as to how I could work around this. Maybe there is something I overlooked?
Thanks
Jeremy
