Hi Everyone,
I am setting up a MPLS L3 VPN Lab and wanted to configure a BGP RR.
In image1, we can see that we are using physical links for each RR router and between them.
Quick question, could it be practical and easier if we put a switch in the middle and use only one subnet? Please advise. If no, please explain why.
It would take away from the redundancy and have your switch set up the switch as a single point of failure.
It can work if you just want to practice configuration.
i did a RR video that might help you…
Image #1 is more like real life, to a degree. Image #2 is faster to set up. If all you want to practice is the MPLS portion, then #2 is probably fine, but #1 would demonstrate things like failover. Also, in real life, there’s no inherent need to transit an RR, and you probably wouldn’t have two RR’s with physical links between them. Moving the 10.10.10.16/30 down to PE1-AS1/PE2-AS2 would be more realistic.
You don’t run BGP inside MPLS core so this scenario does not make sense. Well, technically you could, but in real life you don’t. In MPLS core you use a faster protocol like OSPF or IS-IS and you typically run BGP from PE to CE.
That would make a single failure point and eliminate redundancy.
Good point, forgot about redundancy.
I am running OSPF + LDP in the MPLS core.
But when adding more and more PE routers, we have to full-mesh due to loop prevention mechanism so selecting one router as VPNv4 RR is easier in terms of configuration, and other things.
