BGP Additional Paths

Hi Zain,

You are right, that doesn’t really help :smile: I just fixed this.


Hi Rene,
I have learnt lots of networking stuff in your plain way . God bless you . carry on dear .Thanks


1 Like

Hi Rene,

For eBGP additional paths is working by default, is this correct ?


Hi Fabio,

I’m afraid not. eBGP only advertises a single best path to its neighbors.


Dear Rene,
I am confused about this stuff:

  • neighbor neighbor-id additional-paths send/receive
  • bgp additional-paths select all
  • bgp additional-paths install
  • neighbor neighbor-id advertise additional-paths all

Could you please give me a short explanation?


Hello Thejohn,

  • neighbor neighbor-id additional-paths send: We use this to configure the router so it sends multiple BGP paths to a neighbor.
  • neighbor neighbor-id additional-paths receive: If you have a neighbor that sends multiple paths, that’s nice but you still have to configure your local router that it wants to receive multiple paths.
  • bgp additional-paths select : you receive a bunch of paths from your neighbor but you can still configure your router which of these paths you actually want to use.
  • bgp additional-paths install: this tells the router to actually install a backup path that you selected with the “bgp additional-paths install” command.
  • neighbor neighbor-id advertise additional-paths: This configures your router which additional-paths you want to advertise to a neighbor. “all” means all additional-paths.

These options make more sense if you see them in action. Try to recreate the topology I did in my lesson. When you try each command, look at the BGP and CEF table to see what is actually advertised/received/installed on each device. That makes it much easier to understand all of this.

Hope this helps!


Dear Rene,
If you want to send addtional-path for your neighbor:

  • First, you have to configure additional-paths send feature to send to your neighbor
  • Second, you chose which additional-paths which you want to advertise to the neighbor
    And your neighbor:
  • First, he has to configure additional-paths receive feature to receive from you
  • Second, he has to configure bgp additional-paths install to add which BGP path into his routing table
    I am right? Thanks

Hello Thejohn

Yes, that is absolutely right!


hi Rene and staff,
thanks for always replying to my posts

Just a simple question about what happens just behind configuration R1 to R6 before any consideration about additional paths
R4 and R5 learn Net6 ( via eBGP (next-hop=R6) = OK
R5 has another path to Net6 via R4, but R4 has not the symetric path to Net6 via R5
This is surprising because the topology is symetric

I think the reason is:
R2 as the reflector prefers R4 to go to Net6 (that is the best choice from its perspective)
So, to go to Net6, R2 (the reflector) only advertises the path via R4 to his clients R4 and R5: R5 gets this path (so R5 install a path to Net6 via R4 in its BGP table) and R4 gets also this path from R2, but in this path the next-hop is itself => so this is done
This is why, in R4, there is not a path to Net6 via R5
I just want to know if i am right ??

I will test this in a lab, but i suppose if i shut Gi0/1on R6 (or i shut Gi0/3 on R4) BGP will reconverge via R5 (so in this case, i will find in R4 the route to Net6 via R5)

Hello Dominique

That’s what we’re here for! :slight_smile: It’s always a pleasure.

Your explanation is correct. I suggest you try shutting down the ports you mentioned and allow BGP to reconverge. You should get the behaviour that you predicted.

I hope this has been helpful!



When R1 and R3 are configured to receive the additional paths I can see that the “neighbor additional-paths receive” command is under “router bgp 12345” mode. Shouldn’t the command be under “address-family ipv4” mode?

Many thanks,

Hello Staut

In order to keep a consistency across all routers, it would be a good idea to configure this command under the address-family ipv4 mode. However, it would still function with this configuration since IPv4 is being used for both adjacencies and prefixes.

I hope this has been helpful!


1 Like

Hi Team,

I am still a bit confused on the following…

neighbor neighbor-id additional-paths send/receive
neighbor neighbor-id advertise additional-paths

I understand I need both statements to get this working… but in my mind, i always thought “send”
is “advertise”.

Is there any method to picture or think around this?

Hello Jon

The neighbor neighbor-id additional-paths send/receive command will define if additional paths will be sent and/or recieved with that particular neighbor. In essence, you enable the feature in that particular BGP neighbor adjacency.

The neighbor neighbor-id advertise additional-paths all/best/group-best command indicates which paths will be advertised in that neighbor adjacency. All of them? The best and second best paths? Or the group best paths as defined in the lesson?

So the first command enables the feature between the specific routers, while the second command specifies what kind of paths will be shared.

I hope this has been helpful!


Hi Team,

Thank you very much for the replies.
This helps to clear up my confusion totally.

1 Like

Can be some dumb questions and sorry I have not labbed this up yet.

  1. Can we use additional path install command for paths received in bgp table from different neighbors? For example, I get from Neighbor A and B, can I use this feature to have one of them as backup route?
  2. If we enable maximum-path 2 ibgp on the RR (R2) in this case, won’t it advertise both the paths from R4 and R5 to its clients avoiding all this additional configurations?
  3. It is mentioned that this feature is fro iBGP but for the group best, it is selecting paths from different ASes which is eBGP routes , so is it contradicting , Am I understanding something wrong?

I will give it a read once again, but any explanation will be helpful.


Hello Madhu

Yes you can. Actually, this is precisely what the feature is used for. This can be seen in the lesson, where R2 as a route reflector learned two paths to (from R4 and R5) and advertised them both to R1. With the appropriate configuration parameters, this resulted in the installation of a backup path in R1 as can be seen in the below output:

Note here that the “additional paths” and “maximum paths” features are slightly different. You can use the maximum paths feature, but it will function a little differently. Specifically, Cisco states that:

The bgp additional-paths install command will install the type of path that is specified in the bgp additional-paths select command. If the bgp additional-paths select command specifies both keyword options (best-external and backup), the system will install a backup path.

The maximum-paths ebgp andmaximum-paths ibgp commands trigger a multipath computation, and multipaths are automatically installed as primary paths.

On the other hand, the bgp additional-paths install command triggers computation of a backup path or best-external path.

This info is retrieved from this Cisco documentation.

You are correct that the specific paths being mentioned in the groups are from other AS’es, so we’re looking at eBGP. However, a Route Reflector that is in AS4 for example, may have routes to AS1, AS2, and AS3 via eBGP. However, these routes are advertised via iBGP to other routers in AS4. All the next hop routers may be in the same AS, thus these routes would be shared using iBGP, even though they are indeed destinations outside of the AS.

I hope this has been helpful!


Hi Guys
What about the behavior of R2 for the 6.6.6…6/32 destination ? The BGP convergence on R2 will be right away because it has already installed 2 alternatives in its BGP table:?

  • next hop as the best path installed in the RIB
  • next hop in the bgp table but not in the RIB.

Let’s suppose that R4 is dead and only we have
R2 doesn’t need BGP additional feature to configure an install as a backup route to improve BgP convergence?

Am I right?

Hello Rodrigo

As far as R2 is concerned, it has both paths in its routing table, but advertises only one of them. However, R2 will not immediately converge if R4 goes down. The neighbor adjacency will have to be torn down which requires timers to expire before the specific next hop is considered dead. Once that happens, then BGP will reconverge placing as the next hop for that particular route.

If the additional paths feature is not enabled, BGP must reconverge in order for R2 to realize that it must route packets via R5, which will not be immediate, but will take some time.

I hope this has been helpful!



Interesting. thanks!!!

1 Like