Re: mDNS discussion?

Freek,

The Zeroconf group and this list has been dedicated almost entirely to =
getting the IPv4 Link Local address assignment specification done and =
there has been little discussion for a long time of the other topics you =
mention - in general each of the specs you mention probably has it's own =
list (where you will probably see many of the same people).

Other earlier objectives of the zero-config group were dropped either =
because of lack of progress or because they were being covered =
elsewhere.

For comparison of mDNS vs LLMNR or of SLP vs DNS-SD vs UPnP you are best =
looking not just at the specs and the appropriate lists, but at who is =
promoting or actually using them and in what circumstances and trying to =
assess the reasons (both technical and commercial) for that choice.

Philip

----- Original Message -----=20
From: "Freek Dijkstra" <[email protected]>
To: <[email protected]>
Sent: Monday, July 19, 2004 5:04 PM
Subject: mDNS discussion?


> Hi,
>=20
> I got curious into the operating of ZeroConf, so I viewed the =
zeroconf.org
> and multicastdns.org sites, read an article or two, noted the =
controversy
> with LLMNR, and read draft-cheshire-dnsext-multicastdns-04.
>=20
> This list seems pretty quiet (or the archive is broken...). That's too =
bad.
> I'm just an innocent spectator, not really involved, who just want a =
"zero
> configuration" network. I don't care if the solution is called =
ZeroConf,
> mDNS, RendezVous, DNS-SD, LLMNR or whatever.
>=20
> I understand that zeroconf involves 3 issues topics: automatic IP
> allocation, name resolution and service discovery. (I don't understand =
the
> last one, Automatic allocation of Multicast Addresses, yet).
>=20
> The first, Link-local interface configuration seems more or less a =
done deal
> with the 169.254/16 range and the =
draft-ietf-zeroconf-ipv4-linklocal-17.
>=20
> The third, service discovery, seems focussed around DNS-SD, UPnP and =
SLP.
> I'm probably ignore this (my time is unfortunately limited). My first
> impression at least is that UPnP has unnecessary overhead, but I =
really
> can't judge on DNS-SD and SLP.
>=20
> The second, name resolution seems to focus on either LLMNR or mDNS. =
What is
> the current status of mDNS? Should I conclude from the fact that
> draft-cheshire-dnsext-multicastdns-04 is a individual submission (and =
can't
> thus, according to IETF guidelines become a standards track RFC, but =
only
> experimental or informational track), is not an IETF-endorsed =
activity, but
> (probably just as important) a community-driven activity?
>=20
> Two questions: is the above correct? In particular, are mDNS and LLMNR
> compatible, and if not, does mDNS have functions LLMNR not have? =
Pointers
> are welcome (I will read them both, but a short note or pointer is
> appreciated).
>=20
> Second, having just (partially) read =
draft-cheshire-dnsext-multicastdns-04,
> is it useful to give feedback? I found an obvious type (FF0X::FB in =
section
> 25 should read FF02::FB), but I'm willing to give more feedback, if
> appreciated. If so, were to sent it to? To this list, the authors
> individually, or even [email protected]?
>=20
> Kind regards,
> Freek Dijkstra
>=20
> PS: for the curious and interested: I work for the University of =
Amsterdam,
> and have been to IETF 52&53 when I was a student, but have no =
professional
> interest anymore. Just personal curiousity.