The problem dampening solves
A flap is a route that is withdrawn and then advertised again. A bad cable or a router rebooting in a loop can do this repeatedly. Each flap triggers an UPDATE to every neighbour, which recalculates best paths and passes it on. Across a network the size of the internet that churn costs real CPU time and slows convergence for everyone.
On small screens, swipe sideways or focus the diagram and use the arrow keys.
Penalty, half-life and the two limits
Each time a route flaps, the router adds a penalty to it. The penalty is not fixed: it halves every half-life while the route is stable. Two thresholds turn the number into behaviour.
On small screens, swipe sideways or focus the diagram and use the arrow keys.
On small screens, swipe sideways or focus the diagram and use the arrow keys.
Read the curve left to right. Two flaps in the first ten minutes leave the penalty under 2000, so the route is still used. The third flap at minute 15 pushes it just over the suppress limit and the route is suppressed. It stays suppressed while the penalty halves every 15 minutes, and returns when it falls under 750, about 25 minutes later.
Configuring dampening
router bgp 64500
address-family ipv4 unicast
bgp dampeningR1. Enables dampening with the defaults: half-life 15, reuse 750, suppress 2000, maximum suppress 60.
router bgp 64500
address-family ipv4 unicast
bgp dampening 10 1000 3000 40R1. Custom values in the order half-life (minutes), reuse, suppress, maximum suppress time (minutes). Here the route is suppressed less easily and for a shorter time.
On small screens, swipe sideways or focus the diagram and use the arrow keys.
Verifying dampening
R1#show ip bgp dampening dampened-paths BGP table version is 12, local router ID is 192.0.2.254 Status codes: s suppressed, d damped, h history, * valid, > best, i - internal Origin codes: i - IGP, e - EGP, ? - incomplete Network From Reuse Path *d 203.0.113.0/24 192.0.2.1 00:25:20 64501 i
d marks a damped route. The Reuse column is the time left until the penalty falls to the reuse limit. The route is in the table but is not selected as best while it is damped.R1#show ip bgp dampening flap-statistics Network From Flaps Duration Reuse Path *d 203.0.113.0/24 192.0.2.1 3 00:15:42 00:25:20 64501 i h 198.51.100.0/24 192.0.2.1 1 00:02:10 64501 i
h is a history entry: the route is currently withdrawn but R1 remembers its penalty.clear ip bgp dampening
clear ip bgp dampening 203.0.113.0 255.255.255.0Privileged EXEC mode. Clears the penalty and unsuppresses routes, for all prefixes or for one. Use it after you have fixed the cause of the flap.
Why it is used sparingly today
When one real failure happens, the rest of the internet sees a burst of path changes while it searches for alternatives. The same route can be seen to change several times per router even though its origin flapped once. With aggressive settings, those changes alone push a healthy route over the suppress limit, and a stable prefix disappears from part of the internet for up to an hour. That is the opposite of what you wanted.
RFC 7196 revises the recommended values to far less aggressive ones, and operator guidance such as RIPE-580 recommends against dampening for most networks. If you need to limit instability, fix the cause, and use BFD to detect real failures quickly instead of waiting for the hold timer.
Technical reference: RFC 2439, BGP Route Flap Damping and RFC 7196, Making Route Flap Damping Usable. Output above is illustrative.
Check your understanding
A route has a penalty of 1600 and has been stable for 15 minutes (half-life 15). What is its penalty now?
A route is suppressed. Under what condition is it used again?
Does default dampening on IOS XE suppress a flapping route learned from an iBGP neighbour?