As you can see from Topology, the NBMA routers are all running iBGP, while our outlying router are running eBGP – It just got real!

Now I know what you are thinking, “What if the physical interface connected to the Remote BGP peer goes down? What then smart guy? Just lose the Adjacency of that neighbor???”

That is an excellent question! We use loopback interfaces as neighbors as they will never go to a “down” state unless administratively shut down or the router goes offline (which then we have bigger problems).

Now when I say neighbors, I mean the IP address used in the neighbor statement, that the Peer will send / receive updates and Hellos from. I’m still waiting to see the significance of the RID in BGP, but this is for making the actual neighbor interface a loopback running over physical interface(s).

However, lets demonstrate between R1 and R5 what happens using their loopbacks to create an Adjacency:

First I remove the neighbor configs:

R1(config-router)#no neighbor remote-as 500
[Resuming connection 5 to r5 … ]

*Apr  9 21:22:15.767: %BGP-5-ADJCHANGE: neighbor Down Peer closed the session
*Apr  9 21:22:15.767: %BGP_SESSION-5-ADJCHANGE: neighbor IPv4 Unicast topology base removed from session  Peer closed the session
R5#conf t
Enter configuration commands, one per line.  End with CNTL/Z.
R5(config)#router bgp 500
R5(config-router)#no nei remote-as 100

Now I use the loopbacks in place of the interface IP addresses:

R5(config-router)#nei remote-as 100
[Resuming connection 1 to r1 … ]

*Mar 31 00:41:04.368: %BGP-5-ADJCHANGE: neighbor Down Neighbor deleted
R1(config-router)#nei remote-as 500

I waited a bit but nothing is happening there, as we are still missing some commands to get this thing going as can be seen coming up.

One important BGP note on this topic is that BGP does not consider a loopback interface to be directly connected (even if they are on a common subnet), and also requires a static route to reach the remote loopback to form that neighbor relationship, so we will need a static route on both routers so they know how to reach the now configured neighbors by loopback:

R1(config)#ip route
[Resuming connection 5 to r5 … ]

R5(config)#ip route

So I wanted awhile and no Adjacencies were forming, so I did some verification commands for reachability and such, and found an interesting piece of information:
R5(config)#do ping

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to, timeout is 2 seconds:
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/2/4 ms
R5(config)#do sh ip bgp summ
BGP router identifier, local AS number 500
BGP table version is 1, main routing table version 1

Neighbor        V           AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd                4          100        0               0               1             0    0           never         Idle


So even with reachability to the interface and the neighbor statement configured properly, it is sitting in Idle status, which is not what we want happening (even though we now see the Neighbor as R1’s loopback which is a small victory).

The two commands we will need to deploy in addition to static routes, is “ebgp multi-hop” and “update-source loopback” which will then allow the Peering to complete and Idle to hopefully turn over to Established:

R1(config)#router bgp 100
R1(config-router)#neighbor ?
  activate                 Enable the Address Family for this Neighbor
  advertise-map            specify route-map for conditional advertisement
  advertisement-interval   Minimum interval between sending BGP routing updates
  allowas-in               Accept as-path with my AS present in it
  capability               Advertise capability to the peer
  default-originate        Originate default route to this neighbor
  description              Neighbor specific description
  disable-connected-check  One-hop away EBGP peer using loopback address
  distribute-list          Filter updates to/from this neighbor
  dmzlink-bw               Propagate the DMZ link bandwidth
  ebgp-multihop            Allow EBGP neighbors not on directly connected

  fall-over                session fall on peer route lost
  filter-list              Establish BGP filters
  inherit                  Inherit a template
  local-as                 Specify a local-as number
  maximum-prefix           Maximum number of prefixes accepted from this peer
  next-hop-self            Disable the next hop calculation for this neighbor
  next-hop-unchanged       Propagate the iBGP paths’s next hop unchanged for
                           this neighbor
  password                 Set a password
  peer-group               Member of the peer-group
  prefix-list              Filter updates to/from this neighbor
  remote-as                Specify a BGP neighbor
  remove-private-as        Remove private AS number from outbound updates
  route-map                Apply route map to neighbor
  route-reflector-client   Configure a neighbor as Route Reflector client
  send-community           Send Community attribute to this neighbor
  shutdown                 Administratively shut down this neighbor
  soft-reconfiguration     Per neighbor soft reconfiguration
  timers                   BGP per neighbor timers
  translate-update         Translate Update to MBGP format
  transport                Transport options
  ttl-security             BGP ttl security check
  unsuppress-map           Route-map to selectively unsuppress suppressed
  update-source            Source of routing updates
  version                  Set the BGP version to match a neighbor
  weight                   Set default weight for routes from this neighbor

R1(config-router)#neighbor ebgp
R1(config-router)#neighbor ebgp-multihop ?
  <1-255>  maximum hop count

R1(config-router)#neighbor ebgp-multihop 2
[Resuming connection 5 to r5 … ]

R5(config)#router bgp 500
R5(config-router)#neighbor ebgp-multihop 2
R5(config-router)#do sh ip bgp summ

BGP router identifier, local AS number 500
BGP table version is 1, main routing table version 1

Neighbor        V           AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd         4          100       0       0        1    0    0 never    Idle

So we are still hanging idle there, however I left that monster output to highlight the second command we are missing, “update-source”. Now this is what tells the router which interface the neighbor will get its updates from, so we need to specify the loopback interface being used which will be the loopback local to the router this command is being configured on :

R5(config-router)#nei update-source loopback 5
[Resuming connection 1 to r1 … ]

*Mar 31 01:36:02.675: %BGP-5-ADJCHANGE: neighbor Up
R1(config-router)#nei upd
R1(config-router)#nei update-source loo
R1(config-router)#nei update-source loopback 1

So as can be seen, the loopback needs to be the local loopback as the source to the remote loopback address for it to take, I got a few errors playing with this. Another good thing to note, is that R5 showed the BGP Neighbor status back as up before I could even configure it on R5.

Now for some verification:

Neighbor        V    AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd         4   500       5       5        1    0    0 00:01:36        0    4   100      83      83        1    0    0 01:19:55        0    4   100     126     126        1    0    0 02:03:58        0
[Resuming connection 5 to r5 … ]

*Apr  9 22:17:14.551: %BGP-5-ADJCHANGE: neighbor Up
R5(config-router)#do sh ip bgp summ
BGP router identifier, local AS number 500
BGP table version is 1, main routing table version 1

Neighbor        V           AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd         4          100       5       6        1    0    0 00:01:59        0

R5(config-router)#do sh ip bgp nei
BGP neighbor is,  remote AS 100, external link
  BGP version 4, remote router ID
  BGP state = Established, up for 00:06:15

This is an oddity to me that the 0 in State seems to indicate an Established connection, while the other states will show “Idle” or “Active” while they are waiting to Peer but once Peered it just shows a 0. Very odd.

I also highlighted the “uptime” for the Adjacency between R1 and R3, so now I get a ticket timer to remind me how long I’ve been staring at the screen for while studying, bleh.

With that, I will wr mem around the network, and next up is Route Advertisement with BGP!