ietf-mxcomp
[Top] [All Lists]

Re: FYI - RFC2915 (NAPTR)

2004-05-27 02:53:30


Sorry for yet another post on same subject, but it seems you need
to read this in the following order in the latest RFCs series:
 ftp://ftp.rfc-editor.org/in-notes/rfc3401.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc3402.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc3403.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc3404.txt
as well as possibly these for URN:
 ftp://ftp.rfc-editor.org/in-notes/rfc3405.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc3406.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc2141.txt

On Thu, 27 May 2004, william(at)elan.net wrote:


For those who were interested in NAPTR, going through more DNS related RFCs
I found that 2915 has already been obsoluted by 3403 (at the same time
2915 was already next version and first one was specified in RFC2168).
In any case, this did not seriously change semantics of the record, the 
only thing I can see in 3403 is that they no longer specify protocol flags
and say that these have meaning in application (which is even better
for us). Hers are all these documents if you want to compare:
 ftp://ftp.rfc-editor.org/in-notes/rfc3403.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc2915.txt
 ftp://ftp.rfc-editor.org/in-notes/rfc2168.txt

Since these documents list Verisign people as responsible for their creation
and we seem to have quite a number of Verisign people on this list, perhaps
somebody can comment on why they were introduced and if it seems acceptable
to reuse them for email security application given that we can define marid 
URN or something like that.

And unless somebody says otherwise I'll try to work on the draft myself
that tries to specify how to use NAPTR for MARID application. Please do 
tell me if this is completely bad idea to consider, because right now it 
seems like a perfect use for this record.


P.S. Going through DNS RFCs is quite fun, I find lots of interesting 
records that people thought of for possible future use but it seems it 
never got to it and most of these records I've never seen anywhere. Some
DNS RFCs also seem to contradict each other in some places, so implementing
fully complaint DNS resolver or server seems close to impossible. For 
those interested in email security I would recommend looking at RFC2535, 
RFC2538, RFC3445. RFC2535 actually specified how to enter keys for emails
(aka yahoo domainkeys like), although 3445 said the record should not be 
used this way. Still I think its better to use this or similar scheme then 
to put public key into TXT record like yahoo proposed. Sorry for OT.


<Prev in Thread] Current Thread [Next in Thread>