Ian Jackson wrote:
Brian E Carpenter writes ("Re: Last Call: 'Linklocal Multicast Name Resolution (LLMNR)' to Proposed Standard"):
Russ Allbery wrote:
...
I think your criteria doesn't survive logical scrutiny. If other people
have access to the standard, can implement the standard, and can build on
the standard to create a newer revision of it, I can't imagine what
definition of "proprietary" you're using that would apply.
But mDNS is not on the standards track and LLMNR is proposed to be
on the standards track. That, I think, is why Stuart has raised the issue.
I'm finding this discussion quite disturbing. It seems that the
proposal is that the IETF should bless LLMNR because LLMNR is on the
Blessing Track.
It's been duly submitted by a WG and Last Called, so we *must*
consider it for the standards track. I can't tell you what the result
of that consideration will be; my crystal ball is down.
Surely the reasons for the IETF to bless LLMNR as opposed to mDNS
should be based on technical details
Yes, except that we don't "bless", we approve publication (or not).
and deployment experience ?
Strictly speaking, no - not for Proposed Standard status. But
implementation and deployment experience is always valuable input.
In
which case it seems clear that mDNS is far superior. It's more widely
deployed, more widely implemented, and not COMPLETELY INSANE !
Sorry to be pejorative, but as a DNS implementor[1] I'm amazed to find
senior IETF/IESG people seriously contemplating the kind of namespace
confusion which is fundamental in the LLMNR protocol design.
Can you spell that out please? Since it uses a different port number,
where does the confusion occur?
Brian
The mDNS approach at least isolates the damage (if you consider it
damage) to a specific subregion of the DNS, which you can choose to
have on your search path or in your configuration - or not.
It's a shame they chose `.local' (which was previously thought to be
reserved for local system administrators) but that's too late to
change now. This mistake has happened in part because the relevant
IETF WG failed to properly engage with the mDNS authors !
I'm in general not a big fan of zeroconf, multicast, or anything else
you might think of as weirdo DNS extensions. I'm something of a
luddite. But if we're going to have an IETF-standardised protocol to
address this problem area, surely we should choose the protocol that's
(a) widely deployed, (b) technically superior and (c) less weird ?
I haven't read the drafts in detail, but I have read the LLMNR FAQ,
expecting to find a biased account of the differences which would lead
me to want to read the other side of the story. But in nearly all of
the cases, I found myself disagreeing with the LLMNR way of doing
things, despite the best efforts of the text I was reading !
As long as they are Internet-Drafts they all have the same status, work
in progress, except that LLMNR has gained WG consensus.
From what I can see it might be more accurate to say that the DNSEXT
WG has been captured by people who have a bee in their bonnet about
killing mDNS, or who are at the very least badly misguided.
Ian.
(not usually a fan of people making wild-sounding statements about
IETF WG conspiracies, either!)
[1] GNU adns, http://www.chiark.greenend.org.uk/~ian/adns/
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf