About 20 people attend the gathering Tuesday.
We started off by going over different scenario's in relationships between
"parent" and child DNS Operator
http://dl.dropbox.com/u/81151626/CDS%20Scenarios.pptx
Then followed lively discussion about alternatives and applicability in
different situations including
- child wants to roll KSK algorithm
- child does not want to expose backup KSK but advertise a DS for it
- some child servers are lame which one should parent agent believe
- DNS operator transfer
Among the issues discussed from the child side:
how can a child cause faster uptake of CDS than the polling frequency
how can child learn why CDS has not been accepted
rules for child tools that are performing KSK rollover and what check
to perform
At the end there seemed to be rough consensus that CDS would work in large
fraction of the usage cases
but there are situations where other ways are needed to affect the change of
the DS in parent.
A number of the situations mentioned above will require next version of the
document to include additional
text explaining how/if CDS applies in each situation.
There was some skepticism about the uptake of CDS by certain communities.
thanks everyone that participated in the constructive discussions
Olafur
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
