Link to home
Start Free TrialLog in
Avatar of jskfan
jskfanFlag for Cyprus

asked on

IBGP Routers in Transit AS

IBGP Routers in Transit AS

in the topology below, we have 3 Routers in AS100, I would like to know in Cases:

Case 1:
Router E does not run BGP, then do we have to run IGP(OSPF,EIGRP,RIP) on all 3 Routers [Router A,E,B]  in order to have Router B able to forward EBGP traffic , that has been learned by Router A from AS 300 ,  to AS400 ?
In case we'll have to run IGP in all 3 Routers , then do we need to redistribute BGP into IGP on Router A ?


Case 2: Router E runs BGP as well as Routers A and B, in this case do we still need to run IGP on all 3 Routers in AS 100  or it is not needed?

Thank you



User generated image
Avatar of Soulja
Soulja
Flag of United States of America image

Case 1:
Router E does not run BGP, then do we have to run IGP(OSPF,EIGRP,RIP) on all 3 Routers [Router A,E,B]  in order to have Router B able to forward EBGP traffic , that has been learned by Router A from AS 300 ,  to AS400 ?
In case we'll have to run IGP in all 3 Routers , then do we need to redistribute BGP into IGP on Router A ?


You will need some type of routing. Doesn't have to dynamic. Could be static routes. This is in order for Router A and B to become IBGP neighbors


Case 2: Router E runs BGP as well as Routers A and B, in this case do we still need to run IGP on all 3 Routers in AS 100  or it is not needed?

In this case you'd either need to full mesh the IBGP peerings or use router E as the route reflector. Then just redistribute the direct connected routes into BGP or run an IGP.
Avatar of jskfan

ASKER

Thank you..

In case we'll have to run IGP (OSPF,RIP,EIGRP) in all 3 Routers in AS100 , then do we need to redistribute BGP into IGP from Router A  , or just configure BGP in Parallel with IGP... I mean on each Router  in AS100, we'll configure IGP as well as BGP  ?
If you run an IGP, as long as the IGP has a route to the next hops of the ebgp neighbors you are fine. If not, you will need to add next hop self between the ibgp peers to be able to get to the ebgp prefixes.
Avatar of jskfan

ASKER

Next-Hop-Self  only on RA and RB

However I believe RA,RE, RB  still have to be configured with BGP as well as IGP (OSPF,EIGRP,RIP)

if we Configure it just with BGP only , then we 'll fall on the requirement of  having all 3 routers to be in Full Mesh
My previous comment was based on the assumption of IBGP between A,E, and B.
Full mesh isn't needed if you use E as a route reflector in AS100
Avatar of jskfan

ASKER

By the way , I have seen similar examples like that, where in Routers RA,RE,RB, the BGP peering is done on Loopback interfaces instead of on Physical interfaces
I mean they create Loopback address on RA,RE,RB , and they configure  for instance:
RA# Neighboor 2.2.2.2 remote-as  100
RA# Neighbor 2.2.2.2 Update-Source Loopback 0

I am not sure why they do not build their peering on physical interfaces, seeing that there is only one link to each Router and no other backup links
I meant, I do not understand for instance why they do not configure:

RA# Neighboor  150.10.10.10 remote-as 100
Avatar of jskfan

ASKER

So with Route Reflector  on Router RE, you need to configure BGP on RA,RE,RB , and you do not need IGP (OSPF,EIGRP,RIP)  ?
It's just best practice. Say for example if you do add another path between the routers. You won't have to take down the BGP peering or add an additional one.  Another example is like port channels. If you only have one cable between the switches. I still add it to a port channel just in case in the future I add another.
The only purpose of route reflector is to get around the full mesh requirement.
ASKER CERTIFIED SOLUTION
Avatar of Soulja
Soulja
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of jskfan

ASKER

Thank you N. Spears