0

Consider a network topology like -

                     +-------------------------------------------------+
+--------------------|-----------------------------+                   |
|                    |                             |                   |
|                    |                             |                   |
|                    |                             |                   |
|                    |                             |                   |
|                    |                             |                   |
|   +----------+     |        +----------+         |     +---------+   |
|   |          |     |  BLUE  |          |   RED   |     |         |   |
|   |    R1    |-----|--------|    R2    |---------|-----|   R3    |   |
|   |          |     |  Gi0/0 |          |  Gi0/1  |     |         |   |
|   +----------+     |        +----------+         |     +---------+   |
|                    |                             |                   |
|                    |                             |                   |
|                    |                             |                   |
|                    |                             |                   |
|    BGP protocol    |                             |   EIGRP Protocol  |
|                    |                             |                   |
+--------------------|-----------------------------+                   |
                     +-------------------------------------------------+
                                                                                                                                                                                                                                                                        

Here I am running iBGP protocol between R1 and R2. And I am running EIGRP protocol between R2 and R3.

On R2, the intf Gi0/0, which is connecting to R1 is put under the vrf - BLUE. On R2, the intf Gi0/1, which is connecting to R2 is put under the vrf - RED.

Given that Loopback address, 1.1.1.1, on R1 is advertised onto R2 through iBGP. The route 1.1.1.1 would be stored on the BLUE VRF Routing Table on R2. Now, if we have to advertise the route 1.1.1.1 onto R3, then we will have to first redistribute the route 1.1.1.1 from BGP to EIGRP on R2 from VRF BLUE to RED.

ie,

R2#show run | s router eigrp
router eigrp test                                          
!                                                          
address-family ipv4 unicast vrf RED autonomous-system 100  
 !                                                         
 topology base                                             
  redistribute vrf BLUE bgp 1 metric 1500 100 255 1 1500 
 exit-af-topology                                          
 network 20.0.0.1 0.0.0.0                                  
exit-address-family      

And then EIGRP protocol would take care of advertising it onto R3.

However, we also have to do route-replication from VRF BLUE to VRF RED on R2 like -

R2#show run | s vrf 
vrf definition RED
 !                                                     
 address-family ipv4                                   
  route-replicate from vrf BLUE unicast bgp 1        
 exit-address-family                                   

Why is this additional step of "route-replication" is needed? I know that route-replication would replicate the route from one VRF (BLUE) routing table to another VRF (RED) routing table. If you see the above redistribute command under EIGRP, redistribute vrf BLUE bgp 1 metric 1500 100 255 1 1500 we are essentially telling redistribute all the routes from BGP in the VRF BLUE to EIGRP onto the VRF RED. Then what is the point of replicating the route first from VRF BLUE routing table to VRF RED routing table?

2
  • VRFs are like completely separate routers. They do not share routes unless you explicitly share routes between them. Routes in one VRF cannot be seen by another VRF unless you "create a connection" between the virtual routers.
    – Ron Maupin
    Sep 5, 2022 at 17:10
  • Yes, I get it. As I understand, route-replication is done to replicate the route from one VRF (BLUE) routing table to another VRF (RED) routing table. So after replicating the route, the route should be present in the to VRF (RED) routing table. So why dont we just do intra-vrf redistribution from RED BGP table to RED EIGRP table?
    – Deepak
    Sep 6, 2022 at 15:15

1 Answer 1

2

There are multiple ways to accomplish route leaking. Both redistribute vrf ... and route-replicate will accomplish this.

1
  • My question is, when we try to do redistribution of route from one vrf to another, why should we first do route-replication? Why can't just the command redistribute vrf ... enough? When we do redistribution within the same vrf, we achieve it with the command redistribute ...
    – Deepak
    Sep 6, 2022 at 14:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.