On 29/10/2009, at 9:32 AM, Randy Bush wrote:
folk going down this rathole might consider two things
rpki-rtr suggests that the number of global fetchers will be
radically
lest than the number of global asns
there might be ca chain depth of 3-6 for which a 24 hour cycle would
mean a three day delay, making operators remember curtis unfondly
randy
We modelled producer timing on a particular view of the world. I don't
have a problem that the view of the world has changed, and that
consumer-side timing issues have to be re-considered.
I *think* that even bearing them in mind, the producer side modelling
is still probably valid.
If anyone thinks that producers need to be advised to rendezvous more
than 4x in a 24h period, I think it would help to know. I don't think
we're in crisis point for it going up btw, but I do think that the
cost of processing will come into the limits on this at some point.
In any case, I think abstracting this into operational practice
statements is better than wedging it into architecture.
There is an aspect of the times for re-publication which are published
in certificates and manifests and CRLS defining the lower bound:
everyone has to go fetch inside a cycle time which will respect them,
or wear the consequences for things falling out-of-age.
But the upper bound sanity check, the 'avoid excess work' isn't so
clear to me right now.
-G
_______________________________________________
sidr mailing list
sidr@ietf.org
https://www.ietf.org/mailman/listinfo/sidr