Christian Panigl, ACOnet/UniVie wrote:
Dear Routing WG,
and I'm now asking all participants to come back with comments.
In the observed case all /24 customer networks were cut off from parts of the Internet for more than 2 hours and were no longer able to reach for instance the root nameservers. By the way, many, even top- and second-level nameservers are sitting in /24 (192/TWD) prefixes themselves and could easily be "victims" of such a progressive dampening policy !
We do not implement any dampening on the route-server networks nor some particular networks of strategic importance. This is easily done using a 'deny x.x.x.x in any access list applied to the BGP dampening. This seems like an obvious way to protect from cutting *ourselves* off from important resources.
CP wasn't branding route flap dampening itself, but the aggressiveness of some of the implemented "progressive" parameters and was questioning the real usefulness of progressive dampening at all.
If it encourages sensible aggregation surely this is a good thing ?
* Does flapping really depend on the prefix length?
- To the knowledge of people attending the BOF session no measurements exist. Although several items were already measured by Merit on the stability of routes (as seen in the presentation by G.Winters in the Routing WG) they did not include a stability analysis with regard to the prefix length. If flapping does not necessarily depend on the prefix length longer prefixes should not be punished by more aggressive dampening.
While I have no statistics to back up, personal experiences show that the prefix *is* related to the flap propensity.
* Which networks or prefixes are "important"?
- Long prefixes need not be instable. There are discussions to use long prefix routes ("golden networks") for root nameservers or for other Internet structure servers (even for application servers as news, etc). It can be well assumed that these routes are more stable than others and they must not be dampened too aggressively in order not to tackle the functionality of the Internet itself.
This is easily avoided - see above comments re filters.:)
parameters need to be coordinated throughout the Internet in order to
- allow efficient dampening and easy clearing after repair
- dampen flaps at their source by keeping them from spreading in the network
- flat (default) dampening: because it's very hard to make a distinction between less and more "important", not to say "golden" networks, all prefixes should be treated equally. Efforts should be focussed on the propagation of dampening throughout the Internet.
should come back to the Routing WG list with hints and pointers !
AS1849 (UUNET UK) uses the parameters shown at: ftp://ftp.ripe.net/ripe/presentations/ripe-m25-tbarber-bgp-damp.html These are (more or less) the config used by us in the last few months and have proved to work happily. We have had *very* few enquiries about loss of connectivity and our network has been very very stable (not implying it is because of dampening of course ;-). We have for some time been thinking about changing the policies such that they fall in line with accepted Regional registry policies once these are all alligned. I.e it would be nice to apply zero or Minimal penalisation to any /19 or shorter. This means that any well aggregated route will be less affected that ones which are not. This still leaves us with the legacy of PI space, 192, Holes and multi-homed site prefixes. If the community really does want to move away from such legacies perhaps this kind of wide reaching co-operative action will be a good motivation? If RIPE-routing-wg can come up with a best common practice paper, would most members fall in line ??? MY guess is they probably would. One last comment: Dampening beats the hell out of filtering > /19 ! Regards --Tony