[Gen-art] review of draft-cheshire-dnsext-dns-sd-07.txt

[Francis Dupont <[e-mail]>]

I asked Stephane to review this document...

------------------------------------------------------------------------

I am the assigned Gen-ART reviewer for this draft. For background on
Gen-ART, please see the FAQ at
<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Please resolve these comments along with any other Last Call comments
you may receive.

Document: draft-cheshire-dnsext-dns-sd-07.txt
Reviewer: Stephane Bortzmeyer <[e-mail]>
Review Date: 2010-12-06
IETF LC End Date: 2010-11-23
IESG Telechat date: Unknown


Summary:

The draft is long, containing a lot of interesting material, but not
always well organized. As a general rule, I am a big fan of
"rationale" sections, which let you know why a MUST or a SHOULD was
used. But, here, they are intermixed with the normative content and it
is often difficult to know when standard stops and background material
begins. Examples are sections 4.4 TODO.

In the same way, user interface discussions are very interesting and
useful but should be more clearly separated (may be in a distinct
document).

It also raises a major issue when you put it in perspective with other
DNS-related documents: non-standard terminology and strange ways to
describe the rest of the DNS world. It does not seem suitable for a
Standards-Track document.

Major issues:

There are non-standard terms which may create confusion with existing
technologies. My main concern is with "unicast DNS". There is no such
thing as unicast DNS, there is the DNS and there are other protocols
(like LLMNR or Apple Bonjour name protocol) which are not DNS. Even
worse is "unicast DNS domain name" (unicast refers to a name
resolution method, but a name exists independently of a resolution
method).

Using "local", outside of the procedures of RFC 2606 and 2826 and
similar is a bad idea. I do not elaborate here since the problem is
actually in draft-cheshire-dnsext-multicastdns-12.txt.

TODO: SRV _udp (section 7)

TODO: IANA secret registrations


Minor issues:

For name resolution on the local network, only mDNS (a misnomer since
it is not DNS) is mentioned, not LLMNR (RFC 4795) which seems it would
work as well.

Section 4.1 says that the name must be in NFC. Why not referring to
its subset defined in RFC 5198? Among the practical differences are
the fact that RFC 5198 handles the case of end-of-lines, which is
ignored by the draft.

"commonly-used 16-bit Unicode space" is a misnomer. If one wants to
talk about the Basic Multilingual Plane (which stores only a minority
of Unicode code points), it should use the standard term
<http://unicode.org/glossary/>.

The "standard DNS message size" is no longer 512 bytes since RFC 2671
(section 6.1). Yes, EDNS0 sometimes is blocked by broken middleboxes
but for a technique used only in local networks, this is not a too
serious issue.

TODO: several TXT ina  RRset

The advice of section 13.1 to stuff several other records (which may
be in different bailiwicks) in the DNS answer is at odds with security
practices (RFC 2181, section 5.4.1 and RFC 5452, section 6).


Nits/editorial comments:

Long quotes of existing RFCs (section 4.1) are unecessary since the
other RFCs are available elsewhere.

------------------------------------------------------------------------

IMHO the summary keyword is near to a "Not Ready". BTW there are
unresolved comments from a previous review and of course last call
comments...

Personal comment: a document describing a defacto standard should
not include "rationale" in the standardizing part.

Regards

[e-mail]