Re: [dnsext] one algorithm number or two

[Peter Koch <pk@DENIC.DE>]

On Thu, Dec 11, 2008 at 06:11:19PM -0500, Edward Lewis wrote:

> The question is, how confident are we that NSEC3 is "perfect" and 
> will not be supplanted by something else, in the same manner that 
> NSEC3 has supplanted NSEC (in many people's eyes)?

Ed, I don't think this is a fair question. As we've learned the signalling of
NSEC3 in RFC 5155 actually means that either NSEC or NSEC3 could be used
within the zone.  So, the resolver doesn't know in advance which method
it will see, it is just told to expect either one.  An NSEC3-agnostic
validator will likely treat the zone as insecure. An NSEC3-aware validator
will find its way. Of course, there could be a third flavor that recognizes
the signalling and the NSEC3 RR, but doesn't implement NSEC3 validation.

The same holds for the sha256 aware validator, except that it won't know
for sure in advance to treat the zone as insecure if it doesn't implement NSEC3.
The one thing that is asked is that the resolver recognize the NSEC3 RR type.
The detail that needs to be hashed out, though, is how much "recognition"
is necessary.  Is it sufficient to have "a" (validated) NSEC3 RR in the response
to conclude that there's not going to be an NSEC RR (so as to prove the
NSEC RR(s) ha{s,ve}n't been stripped of) or does the validator actually have
to do the hashing -- which would probably half way of implementing NSEC3
already?

This is independent of general deployment and independent of the actual use
of NSEC3 in a particular zone.  I'd assume there will be a number of zones
where NSEC3 doesn't buy much and which will continue to use NSEC anyway.

-Peter

--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>