eBGP with loopback address question

[Jean-Francois Gagnon]

Hello!

I was wondering if I followed good practices and if my config make any sense.

So, I have a 4 routers in eBGP configured and advertising networks inside.

R1--------------------------

interface Loopback0
ip address 1.1.1.1 255.255.255.255
!
interface FastEthernet0/0
ip address 172.16.1.1 255.255.255.0

!

interface FastEthernet2/0
ip address 172.16.2.1 255.255.255.0

router bgp 100
no synchronization
bgp log-neighbor-changes
network 172.16.2.0 mask 255.255.255.0
neighbor 1.1.1.2 remote-as 100
neighbor 1.1.1.2 update-source Loopback0
neighbor 1.1.1.3 remote-as 100
neighbor 1.1.1.3 update-source Loopback0
neighbor 1.1.1.4 remote-as 100
neighbor 1.1.1.4 update-source Loopback0
auto-summary
!

ip route 1.1.1.2 255.255.255.255 172.16.1.2
ip route 1.1.1.3 255.255.255.255 172.16.1.3
ip route 1.1.1.4 255.255.255.255 172.16.1.4
!

R2---------------------

interface Loopback0
ip address 1.1.1.2 255.255.255.255
!
interface FastEthernet0/0
ip address 172.16.1.2 255.255.255.0

interface FastEthernet2/0
ip address 172.16.3.1 255.255.255.0

router bgp 100
no synchronization
bgp log-neighbor-changes
network 172.16.3.0 mask 255.255.255.0
neighbor 1.1.1.1 remote-as 100
neighbor 1.1.1.1 update-source Loopback0
neighbor 1.1.1.3 remote-as 100
neighbor 1.1.1.3 update-source Loopback0
neighbor 1.1.1.4 remote-as 100
neighbor 1.1.1.4 update-source Loopback0
auto-summary

ip route 1.1.1.1 255.255.255.255 172.16.1.1
ip route 1.1.1.3 255.255.255.255 172.16.1.3
ip route 1.1.1.4 255.255.255.255 172.16.1.4

---------------------------------------------------

So I was wondering if it was good practice that the next hop for received routes are the loopback address of the advertising router. Would'nt be better to be the Fa0/0 interface IP address?

Thanks for your help!

[Francesco Molino]

Hi

First of all, can you share a quick design of how your routers are connected together, just to be sure?

Based on your configs, there're all connected through the same subnet on interface Fa0/0.

Using Loopback addresses as update-source is mostly used within iBGP and not eBGP.

However, we use Loopback is eBGP design, when you have multiple path to your eBGP peering and don't want to use your peering if the physical interface goes down. 

In your case, it doesn't seem to be the case and then no need to use Loopback as update source.

Also, when using Loopback as update source interface in eBGP, you'll need multihop activated.

[Julio E. Moisa]

Hi

I agree with Francesco, usually the loobpacks are used for iBGP, if you want to use the loobpacks with eBGP peering you need configure a NLRI to advertise the loopbacks addresses, the NLRI can be any other routing protocol or static routing but take in consideration that BGP cannot advertise the loobpacks because it could generate a flapping affecting the BGP peering due to the administrative distances (for example eBGP 20 and OSPF 110), the best way is using the directly connected IP addressess for eBGP. 

Hope it is useful

:-)

[Jean-Francois Gagnon]

My bad, it is a iBGP. Sorry

[Jean-Francois Gagnon]

My bad, it's really a iBGP.

And so, I was wondering if it was good practice that the next-hop is the loopback ip address of the neibhor or should it be hte ip of the interface

Thanks

Network Next Hop Metric LocPrf Weight Path
*>i172.16.3.0/24 1.1.1.2 0 100 0 i
*>i172.16.4.0/24 1.1.1.3 0 100 0 i
*>i172.16.5.0/24 1.1.1.4 0 100 0 i

[Julio E. Moisa · Accepted solution]

It is more flexible to reach a BGP peer if a main path is down so other path can be used to reach the peer. You can use RR or BGP confederation to avoid full mesh BGP. The first option is more used. 

[Francesco Molino · Accepted solution]

As I said in my first reply, that's fine using Loopback as update-source interface as usually in iBGP you have a full mesh topology.
To have a better design, you can add or elect one of them as Route-Reflector to not face issues due to iBGP loop prevention system.