Hello,question about BGP IGP tag:
On Cisco when i'm doing [aggregate-address summary-only] command the summary route is created by BGP is with IGP tag. Even when no of suppressed routes has an IGP flag.
But the RFC says:
ORIGIN attribute: If at least one route among routes that are
aggregated has ORIGIN with the value INCOMPLETE, then the
aggregated route must have the ORIGIN attribute with the value
INCOMPLETE. Otherwise, if at least one route among routes that are
aggregated has ORIGIN with the value EGP, then the aggregated
route must have the origin attribute with the value EGP. In all
other case the value of the ORIGIN attribute of the aggregated
route is INTERNAL.``
Looking into that i think that Cisco IOS-XE behaves not compliant to RFC because no of the suppressed routes are with IGP tag. Here is my BGP route table with aggregate address command:
aggregate-address 10.254.33.0 255.255.255.0 summary-only
s> 10.254.33.0/31 0.0.0.0 0 32768 ? *> 10.254.33.0/24 0.0.0.0 32768 i s> 10.254.33.2/31 0.0.0.0 0 32768 ? s> 10.254.33.4/32 0.0.0.0 0 32768 ? s> 10.254.33.5/32 10.254.33.2 201 32768 ? s> 10.254.33.6/31 10.254.33.2 200 32768 ? s> 10.254.33.8/31 10.254.33.0 200 32768 ? s> 10.254.33.10/31 10.254.33.0 200 32768 ? s> 10.254.33.12/31 10.254.33.2 200 32768 ? s> 10.254.33.14/32 10.254.33.2 210 32768 ? s> 10.254.33.15/32 10.254.33.2 210 32768 ? s> 10.254.33.16/31 10.254.33.0 200 32768 ? s> 10.254.33.18/31 10.254.33.2 200 32768 ?
Could you say why Cisco creates summary route with IGP tag as the RFC does not say so ?
it doesn't work in pre-XE (15.6) either...
*> 126.96.36.199/24 0.0.0.0 32768 i
s> 188.8.131.52/30 0.0.0.0 0 32768 i
s> 184.108.40.206/30 0.0.0.0 0 32768 i
s> 220.127.116.11/30 0.0.0.0 0 32768 i
s> 18.104.22.168/30 0.0.0.0 0 32768 i
s> 22.214.171.124/30 0.0.0.0 0 32768 ?
Thanks, from your output seems like it's the same behavior. Would be nice to know from Cisco if it's intended BGP behavior
>> Could you say why Cisco creates summary route with IGP tag as the RFC does not say so ?
To make easier the life in multi vendor environments in big service provider networks, because for Example I know for sure that Juniper networks do this, IOS XR do this, I suppose ALU and Nokia do this
if you have a mix of IOS XE and IOS XR PE nodes it is handy too.
Hope to help
Thanks, sad but FRR (soon too be new Linux Routing stack) works by RFC. When u have multiple routes received from FRR box and CISCO box - these routes become different (by IGP tag) and ECMP doesnt work, so i need to use some workarounds like set igp in route-maps.