Re: [dnsext] DNS RRTYPEs for ILNP review - Comments period end July 5th

[Tony Finch <[e-mail]>]

> http://tools.ietf.org/id/draft-irtf-rrg-ilnp-dns-05.txt

This looks OK to me, though I have a few observations.

Regarding the presentation format of NID and L64 RRs, is the uncompressed
NNNN:NNNN:NNNN:NNNN format going to be standard throughout ILNP? I
couldn't find any mention of it in draft-irtf-rrg-ilnp-arch. The DNS
master file presentation format should be the same as the usual syntax in
other contexts.

>          The CNAME record is closest conceptually to an LP
>          record, but a CNAME is a node name referral scheme,
>          while the LP record is indicating that the given node
>          has the same routing prefix as some other domain name,
>          but does not necessarily have any other values that are
>          the same.

Actually the RT "route through" resource record [RFC1183] is exactly the
same syntax and almost exactly the same semantics as the LP record. The
only difference I can see is which resource records the querier expects to
find at the target name, and the corresponding additional section
processing. That might be enough to justify the new RR type.

There are some textual problems with the specification of the LP RDATA:
the text describing the presentation format of the target domain name
actually describes the RDATA wire format.

Tony.
-- 
f.anthony.n.finch  <[e-mail]>  http://dotat.at/
Southeast Biscay: Variable 3, becoming southeasterly 4 or 5 later. Slight or
moderate, occasionally rough later. Occasional rain. Moderate or good.