On Apr 6, 2014, at 6:56 PM, Edward Lewis <[email protected]> wrote:

> 
> On Apr 6, 2014, at 20:52, Paul Hoffman <[email protected]> wrote:
> 
>> That does not make the document "impossible to implement", it makes it 
>> unreliable when the child's nameserver is anycast. That's why the document 
>> says not to do that in Section 2.2.1
> 
> My comment wasn’t that the “document was impossible to implement” but that 
> with the advent of any cast one can never query all of the name servers (from 
> one location).
> 
>> Again: isn't your concern covered by the four steps in Section 2.2.1?
> 
> As far as the SOA and NS, that is a good approximation, but it doesn’t cover 
> the A and AAAA records that could be elsewhere.
> 
> It is not that this stuff can’t work.  It’s that ideas like this (adding to 
> the DNS protocol’s management toolset) are only ever approximations that run 
> the risk of adding complications when things don’t go well.  The alternative 
> is another protocol for this, and I expect a blank stare in return to the 
> comment.  So in the mean time, I just want to call out the rough edges of 
> these proposals.

Calling out rough edges is fine; those should be admitted in the documents in 
question. My review of draft-ietf-dnsop-child-syncronization says that the 
rough edges are called out; if you have particular places in the document that 
you think should have bigger warnings, point them out.

> The document does do a good job of handling as many as it can, but there’s 
> still the underlying limitations of one-way-only message passing.   The only 
> way this will work (in the sense of scale and stability over time) is if the 
> parental agent gets notified reliably to check and the child’s (agent?) stays 
> on the ball until the action is completed.

That is a good summary; maybe it should be added to the end of Section 1.

--Paul Hoffman
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to