To:
bill <bmanning@karoshi.com>
Cc:
Alain Durand <Alain.Durand@Sun.COM>, dnsop@cafax.se, Jun-ichiro itojun Hagino <itojun@itojun.org>
From:
Mans Nilsson <mansaxel@sunet.se>
Date:
Thu, 13 Nov 2003 11:28:52 +0100
Content-Disposition:
inline
In-Reply-To:
<200311121930.hACJUuN02483@karoshi.com>
Sender:
owner-dnsop@cafax.se
User-Agent:
Mutt/1.4.1i
Subject:
Re: DNS discovery
Subject: Re: DNS discovery Date: Wed, Nov 12, 2003 at 11:30:56AM -0800 Quoting bill (bmanning@karoshi.com): > > well, I can see nodes that move between topologies, some > where RA is used/prefered and some where DHCP is used prefered. > historically, the SA's for DHCP service are distinct from the > SA's who manage routing, so moving forward, we are providing > a tool to each community that performs essentially the same > task. The end systems have had to deal with getting IP > addresses assigned from multiple places (static, RA, DHCP) > already - listing DNS server preference when there are > multiple inputs -can't- be that hard. It means more code, more bugs, more interop problems. > And its not like the IETF has not promoted two solutions > in the past... can you say OSPF and ISIS? Sure you can. End nodes (the devices who receive autoconfig benefits) typically do not talk IGP, except that they may discover a route via RA's or get it via DHCPv4. Thus, the choice of IGP does not bother the end nodes. It is more like using vi or emacs to edit zone files. Does not affect the casual user/autoconfig node. So, my point is that the "interface to the masses" needs to be uniform, as uniform as can be. One resolver configuration method (besides $EDITOR /etc/resolv.conf) makes the world a less confused place. And, I think we're in agreement on WKA. They are a can of worms. -- Måns Nilsson Systems Specialist +46 70 681 7204 KTHNOC MN1334-RIPE I fill MY industrial waste containers with old copies of the "WATCHTOWER" and then add HAWAIIAN PUNCH to the top ... They look NICE in the yard ...