@mihaigabriel wrote:
Hello,
I am reading the Junos Mpls and Vpns course and I can't understand what is the practical utilization of l2vpn interworking stitching or l2vpn to l3vpn interconnection.
Could someone enlighten me?
Thanks
Hi Gabriel,
The main reason to perform L2VPN stitching into say an L3VPN is when you want to backhaul a CE that is located in one part of the network to another PE that provides an L3 service.
Normally, when you configure an L3VPN VRF, it is expecting that all the local CEs are reachable on local interfaces on that PE device. What if however, there is a CE that is located in some other portion of the network and terminates on a PE that does not participate in the L3VPN instance or provide such a service. Well, if we can stretch the L2 segment from that CE to the PE that actually provides such an L3VPN service using an L2 pseudowire in the core it will allow us to support CEs that are not directly connected to the PE that is providing the L3 VRF.
This is just one example of one corner case, but there are many others - for example, maybe you want to provide a VPLS service throughout your network, but there is an older legacy PE that does not support VPLS. Well, if it happens to support an L2VPN we might be able to backhaul the L2 circuit to another PE that can actually provide the VPLS service. For this we would need VPLS to L2VPN stitching.
Anyway, hope that helps, and feel free to ask any other questions you might have. BTW, this would probably be a more appropriate question for the Routing Section.
Cheers!