I didn't mean to say it is required. But I think it is so common these days that
it's fair to regard it as normal for large intranets.
Brian
[EMAIL PROTECTED] wrote:
>
> I disagree that we can assume any site will have two face DNS. Thats a bad
> assumption.
> /jim
>
> > -----Original Message-----
> > From: ext Brian E Carpenter [mailto:[EMAIL PROTECTED]]
> > Sent: Wednesday,February 07,2001 10:19 AM
> > To: Robert Elz
> > Cc: Paul Francis; [EMAIL PROTECTED]
> > Subject: Re: another renumbering question
> >
> >
> > Once you *know* that a host has a site local address, the
> > address selection draft tells you when to use it. I've always
> > assumed that they would be in the DNS along with global addresses.
> > Presumably large sites will be running two-faced DNS anyway, so
> > these addresses will never go outside.
> >
> > Brian
> >
> > Robert Elz wrote:
> > >
> > > Date: Tue, 6 Feb 2001 16:24:23 -0800
> > > From: "Paul Francis" <[EMAIL PROTECTED]>
> > > Message-ID: <006b01c0909c$528dda60$1300a8c0@dellchan>
> > >
> > > | But I can't recall how it is a host decides when
> > another host can be reached
> > > | via a site-local address, nor can I find where I read that text.
> > >
> > > This is still one of life's mysteries I think - several ideas have
> > > been floated, but nothing yet committed to a draft that I'm
> > aware of.
> > >
> > > The include placing the site locals in the DNS, and having clients
> > > do a match on the global addresses to see if the site local should
> > > refer to the same site (one that finds no attraction to me at all).
> > >
> > > And having the client send from its site local addr (so the
> > packet cannot
> > > leave the site) to the remote global addr an ICMP saying
> > "tell me your
> > > addresses", and receiving a list that includes the site
> > locals for the
> > > destination (or receiving an ICMP error indicating an
> > attempt to cross
> > > the site boundary with a site local source addr). (If the
> > source doesn't
> > > have, or doesn't want to use, its site local addr, then
> > there's no point
> > > doing any of this, may as well just use global addr of the
> > dest a well).
> > > This one I like - it adds a small delay in communications
> > the first time
> > > a connection is attempted to a new node (only for
> > connections initiated,
> > > responses would never do this) but that is bounded by the
> > RTT to the edge
> > > of the site, which is usually of the order of a couple of ms.
> > >
> > > kre
> > >
> > > --------------------------------------------------------------------
> > > IETF IPng Working Group Mailing List
> > > IPng Home Page: http://playground.sun.com/ipng
> > > FTP archive: ftp://playground.sun.com/pub/ipng
> > > Direct all administrative requests to [EMAIL PROTECTED]
> > > --------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
