BGP Part 3: Advanced Topics
Root Flap Dampening
Each time one of the prefixes flaps, there is gonna be a penalty. Default penalty of 1000 per flap. Default supress limit is 2000. Default reuse limit is 750.
Half-life concept: The penalty will be reduced by half every 15 minutes.
Maximum supress limit - 4 time the half life (60 minutes)
Configuration, will activate the feature with default parameters.
bgp dampening
To configure specific
bgp dampening half-life reuse supress max-supress
show ip bgp flap-statistics
show ip bgp dampened-paths
clear ip bgp flap-statistics
clear ip bgp dampening
Latest recommendation from RIPE:
supress threshold 6000
max supress threshold 50000
Example:
we flap int lo0 in the toplogy, so it will go to the history mark. If we flap it enough times, it will go to the dampening state.
We can check penalty on the prefix
Outbound Route Filters
With ORF, we require a capability command on both the sender and the receiver of prefixes. If R2 is sending to us prefixes and we want onlu one of them ,the usual way to do it was to deny it, but resources/bandwith still would be used. So what if we simply tell R2 that we want from him only 1 prefix. Why send me prefixes I am going to filter anyway?
neighbor {IP_Address} capability orf {prefix-list} [both | receive | send]
show ip bgp neighbor {IP_Address} received prefix-filter
- ORF supports Ipv4 and Ipv6 UNICAST routes
- does NOT support IP multicast routes
- can be used only with prefix-lists (no distribute or acl)
- has to be configured separately for every address family (can not be configured under the global BGP process)
- supported only over eBGP sessions