networks reachable within an mpls network

[jonathan johnson]

Hi all,

I'm hoping I might be able to get some insight into how a BGP free core mpls network handles certain senarios.

I understand that with a BGP free core you can have a customer connected to a PE and announce a network which will be reachable from a PE at the other end of the network - even though the core routers don't have a route for it.

My question is what is if I have non BGP speaking ethernet customer connected to the PE via a sub-interface (for example) that has a network statically routed to them or just uses the assisgned point-to-point address along with NAT to gain connectivity to the internet and/or other resources.  How is this static route or point-to-point address reachable by the PE at the far end?  What mechanism annoucnes the assigned network to the far end PE?

I'm sorry if I'm missing something very fundamental - any assistance would be appreciated.

Thanks,

Jonathan.

[Luc De Ghein]

Hi Jonathan,

The routing needs to be there end-to-end or PE-to-PE, so in overlay. BGP running between the PE routers takes care of that. So, the customer or VPN routes need to be advertised by BGP between the PE routers. So, the IGPs of the VPN, but euqally so the connected routes (PE-CE link) and static routes need to be advertised in this fashion. This is achived by redistributing the IGP, the connected or the static routes into BGP. When running BGP between the PE and CE routers, no redistribution is needed. The BGP routes from the CE are transported by the PE to the remote PE routers automatically.

So, in your case, you'll need to "redistribute connected" or "redistribute static" into BGP on the PE router.

regards,

Luc

[jonathan johnson]

Hi Luc,

Thanks for the reply, it helps a lot with my undertanding of how it should work.

All the best,

Jonathan