> I think the question that I have yet to see answered that Dan is
> not articulating well is "how the hell is DNS glue supposed to work
> sanely with A6?" I can't find an answer to that myself. Everybody's
> talking about no full IP addresses written anywhere,
Whoa! Who is this "everybody"?
> and DNS delegation is one place where that just doesn't work. You
> need the full 128 bit address in there, and glue registry is the
> one place that these addresses can't change frequently, because of
> the bulk of TLD zones and complicated, time consuming procedures
> registrars use to update those records...
I think you aren't intending to make a case in favor of well-
segemented A6 chains, but your ellipsis could be completed in exactly
that way. Then if you the end site choose to renumber, you're
responsible for coordinating the TLD's glue. If renumbering is
thrust upon you because your ISP is renumbering, they are responsible
for it. And if the glue record of concern is the very same one they
use for their own DNS infrastructure, the right sort of motivations
are in place.
But never mind that, because the world isn't ready to play that way
and it isn't what RFC 2874, section 5.1.2 suggests:
5.1.2. Glue
When, as is common, some or all DNS servers for X.EXAMPLE are within
the X.EXAMPLE zone itself, the top-level zone EXAMPLE must carry
enough "glue" information to enable DNS clients to reach those
nameservers. This is true in IPv6 just as in IPv4. However, the A6
record affords the DNS administrator some choices. The glue could be
any of
o a minimal set of A6 records duplicated from the X.EXAMPLE zone,
o a (possibly smaller) set of records which collapse the structure
of that minimal set,
o or a set of A6 records with prefix length zero, giving the entire
global addresses of the servers.
The trade-off is ease of maintenance against robustness. The best
and worst of both may be had together by implementing either the
first or second option together with the third. To illustrate the
...
If the EXAMPLE zone includes redundant glue, for instance the union
of the A6 records of the first and third options, then under normal
circumstances duplicate IPv6 addresses will be derived by DNS
clients. But if provider DNS fails, addresses will still be obtained
from the zero-prefix-length records, while if the EXAMPLE zone lags
behind a renumbering of X.EXAMPLE, half of the addresses obtained by
DNS clients will still be up-to-date.
The zero-prefix-length glue records can of course be automatically
generated and/or checked in practice.
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
