BGP (metric x) vs. IGP metric - how to change the BGP "ttl" (metric x) value

[2colin-cant]

Hi guys,

we are using the additional-path install aka BGP PIC feature on IOS on four routers using the following:

address-family vpnv4
bgp additional-paths install

Diagram of topology:

W1 ---- C1 ----- C2 -----W2 

|                                             |

|                                             |

ISP1                                     ISP2

|                                             |

|                                             |

------------------------------
        192.168.200.1/32

Now I do have the best / repair path visible. However I have noticed the metric in brackets (metric x)

where that metric is a different "hop count" from the IGP metric.

Is there any way to influence that BGP hop count, as I believe it is being calculated by the TCP 179 TTL value to gather this information. (to be confirmed)

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

C1#show ip bgp vpnv4 vrf CUSTOMER-X 192.168.200.1
BGP routing table entry for 2218:9110100:192.168.200.1/32, version 9560
Paths: (2 available, best #2, table CUSTOMER-X)
Additional-path-install
Advertised to update-groups:
300
Refresh Epoch 9
65530, imported path from 2218:9110800:192.168.200.1/32 (global)
172.16.254.8 (metric 3) (via default) from 172.16.254.8 (10.16.254.8)
Origin IGP, metric 0, localpref 150, valid, internal, backup/repair
Community: 911080501
Extended Community: RT:2218:9110000 , recursive-via-host
mpls labels in/out nolabel/40
rx pathid: 0, tx pathid: 0
Refresh Epoch 1
65530, imported path from 2218:9110700:192.168.200.1/32 (global)
172.16.254.7 (metric 2) (via default) from 172.16.254.7 (10.16.254.7)
Origin IGP, metric 0, localpref 150, valid, internal, best
Community: 911070500
Extended Community: RT:2218:9110000 , recursive-via-host
mpls labels in/out nolabel/57
rx pathid: 0, tx pathid: 0x0

Even said the PIC feature works, we would like to prefer ISP1 path solely, while keeping the repair path.

But if I try use Local-Pref / IGP metrics / origin  I loose the repair path out of the BGP table.

Hence I am keen to understand if there is a IOS command documented or not, which I can adjust that "TTL" value being received by the process.

In my case I would like to have 10.16.254.7 preferred over .8

Thanks

Colin   - www.flashcardguy.ch

[Jeroen Huysmans]

Hello,

your red metric shows the metric learned to reach the next-hop that has access to the specific subnet.

In an MPLS network this could eg be the OSPF metric to reach the next-hop... In your example 1 next-hop has a lower metric...

The blue metric reflects the BGP metric for the specific subnet you try to reach. It's the same for both entries in your example.

In the BGP path selection mechanism the choice for your example is made (if I'm correct) by the following step:

Prefer the path with the lowest IGP metric to the BGP next hop.

https://www.cisco.com/c/en/us/support/docs/ip/border-gateway-protocol-bgp/13753-25.html

assuming your diagram, it makes sense that from C1 point of view the left path would be shorter...

I have no knowledge on PIC, but when you take a look at the path selection mechanism Local pref is the 2nd in line to make a routing decision...

regards,

Jeroen