opensubscriber
   Find in this group all groups
 
Unknown more information…

d : dnsop@ietf.org 13 July 2012 • 4:07AM -0400

Re: [DNSOP] draft-wouters-dnsop-secure-update-use-cases-00
by Joe Abley

REPLY TO AUTHOR
 
REPLY TO GROUP





On 2012-07-11, at 11:09, Chris Thompson wrote:

> [That's not to say that any differences between the two NS RRsets is ever
> desirable, except as may be necessary or expedient during a change.]

That's a good rule of thumb, but so long as the union of the delegation and apex NS sets contains no lame delegations, everything will still work as far as the client is concerned.

I have seen someone use a deliberately different apex and delegation NS set (with all servers authoritative) in order to try and identify differences in the traffic pattern (e.g. to see whether there are clients who reliably cache and re-use the delegation NS set in spite of the apex NS set).

When submitting NS set changes to the root zone it is required to update the apex NS set before submitting the change request, and until the change request is completed the delegation set will be definitively different. Other registries that perform configuration checks on child zones might behave differently. I realise this was encapsulated in your use of "necessary" but I thought perhaps it was worth spelling out.


Joe
_______________________________________________
DNSOP mailing list
DNSOP@ietf...
https://www.ietf.org/mailman/listinfo/dnsop

Bookmark with:

Delicious   Digg   reddit   Facebook   StumbleUpon

Related Messages

opensubscriber is not affiliated with the authors of this message nor responsible for its content.