Randy Bush [EMAIL PROTECTED] writes:
as fred succintly said, how often does renumbering occur? how often do
lookups occur? so, for which should we optimize?
I can imagine some mobile devices/subnets that are renumbered more
often than their address is looked up. But, if there's only
as fred succintly said, how often does renumbering occur? how often do
lookups occur? so, for which should we optimize?
I can imagine some mobile devices/subnets that are renumbered more
often than their address is looked up. But, if there's only one nice
thing about A6, it's that you get
The issue is not the vintage of the resolver -- it is a UDP datagram
length problem. We assume a much larger v6 datagram can get through
without the risk of fragmentation.
I can't believe the consideration that led to the nonsensical limit
of 512 was the risk of fragmentation in transit. I'm
I think it was recognized a long time ago that the initial deployment
of A6 records should be limited to two (or at most 3) levels. ...
I see... it is unfortunate RFC2874 does not cover it.
it has A6 reference from leaf customer A6 record to ISP A6 record,
and 5 levels of
Date:Wed, 24 Jan 2001 09:03:03 -0500
From:Keith Moore [EMAIL PROTECTED]
Message-ID: [EMAIL PROTECTED]
| What would be rational is for the complete address to be updated
| at the primary server for the zone and propagated from there to
| secondaries, caches,
Date:Tue, 23 Jan 2001 12:15:31 -0600
From:[EMAIL PROTECTED]
Message-ID: B9CFA6CE8FFDD211A1FB0008C7894E4603063B3F@bseis01nok
| Those debates were before running code. The running code is not going to
| scale and is not scaling. Or does that not mean anything
as fred succintly said, how often does renumbering occur? how often do
lookups occur? so, for which should we optimize?
I can imagine some mobile devices/subnets that are renumbered more
often than their address is looked up. But, if there's only one nice
thing about A6, it's that you get
%
% Has anyone supplied an IPv6 address for their name server when
% registering or updating a domain directly under COM, NET, ORG or EDU?
% If so, what happened?
%
% VeriSign Global Registry Services' infrastructure does not (yet) support
% IPv6 addresses: your registrar can't specify an
"Cricket Liu" [EMAIL PROTECTED] writes:
Regarding this proposed "reply with A6/ RRs only if queried
over IPv6" suggestion: Even if we imagine this enhancement patched
in to BIND 9.1.0 and deployed on the root name servers, how are we
going to prevent non-root name servers from
Matt Crawford wrote:
Has anyone supplied an IPv6 address for their name server when
registering or updating a domain directly under COM, NET, ORG or EDU?
If so, what happened?
The result is ...
We are unable to process the Host Form below because
of an error within the form. In
Date:Fri, 19 Jan 2001 11:35:14 -0500 (EST)
From:Jim Bound [EMAIL PROTECTED]
Message-ID: [EMAIL PROTECTED]
| A6 is PS not DS. If it will not work as is we have every right to oppose
| it.
All of IPv6 is DS (or less), not full std, so the same could be said.
|
Has anyone supplied an IPv6 address for their name server when
registering or updating a domain directly under COM, NET, ORG or EDU?
If so, what happened?
-
The IPv6 Users Mailing List
Unsubscribe by sending "unsubscribe
| I also think the
| processing of A6 should not be done on clients but on the servers.
That argument was made (by you, and others) back when A6 was being
debated. If you care to look back over the archives (IPNG, and
perhaps DNSIND) you will see why things are the way they are.
I'm sure
Eidnes; [EMAIL PROTECTED]; [EMAIL PROTECTED];
[EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: Re: (ngtrans) Re: IPv6 dns
Randy Bush [EMAIL PROTECTED] writes:
this has implications for the roots; but perhaps also, for their
ISPs and for address assignment in general. regardless of what
assumptions
Date:Sat, 20 Jan 2001 21:13:22 +0900
From:[EMAIL PROTECTED]
Message-ID: [EMAIL PROTECTED]
| could you give us more concrete example, so that we can share the
| image in your mind?
|
| (i am 100% guessing here) if you mean something like dialup
nothing different from v4, no?
fundamentally, no. in either case you want to arrange for the root
servers to have addresses that are more stable than the average IPv#
address.
Keith
-
The IPv6 Users Mailing List
%
% Perry,
%
% regarding the new root servers which respond to queries over IPv6 -
% one thing you didn't quite mention, but I suspect is necessary, is
% that their IPv6 addresses be stable. that is, that once root
% servers are established at these addresses and are advertised,
% that there
% It has occurred to me for some time that we should be willing to
% supply host routes for the root servers in the default free
% zone. We're only talking about a tiny number of servers, and it is the
% only service for which this is necessary. Yes, it would end up adding
% a dozen extra routes
Bill Manning [EMAIL PROTECTED] writes:
THe last time it was seriously raised was at the Joint IETF/ISOC mtg in
Montreal. The failure modes are pretty spectactular, at least until
DNSsec is deployed and applications can verify the accuracy of the data
received from a root server.
You can
nothing different from v4, no?
Marc.
At/ 21:18 2001-01-19 -0500, Keith Moore you wrote/vous criviez:
Perry,
regarding the new root servers which respond to queries over IPv6 -
one thing you didn't quite mention, but I suspect is necessary, is
that their IPv6 addresses be stable. that is, that
Marc Blanchet [EMAIL PROTECTED] writes:
nothing different from v4, no?
I think the assumption has been in the v6 world that we'll be
renumbering a lot more often. There are, however, a few hosts that we
pretty obviously don't know how to renumber often: the root name
servers.
Perry
Bound [mailto:[EMAIL PROTECTED]]
Sent: Friday, January 19, 2001 8:35 AM
To: Perry E. Metzger
Cc: [EMAIL PROTECTED]; Christian Huitema; Randy Bush; Bill Manning;
[EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: Re: (ngtrans) Re: IPv6 dns
A6 is PS not DS. If it will not work as is w
Randy Bush [EMAIL PROTECTED] writes:
this has implications for the roots; but perhaps also, for their
ISPs and for address assignment in general. regardless of what
assumptions we make about IPv6 renumbering elsewhere, we don't
want to renumber the root servers very often.
and i
| as fred succintly said, how often does renumbering occur? how often do
| lookups occur? so, for which should we optimize?
Depends what the user of the number, surely? There are many (current
IPv4, and so one would presume, quite possibly IPv6 as well) addresses
that are changed much
I agree with you we should be fine deploying AAA records. I also agree
we need to be skeptical about A6 and if they continue to exist we need to
for sure do the type of experiments Randy suggested.
mornin' jim,
to be pedantic, i did not suggest a particular experiment. what i asked was
Jim Bound [EMAIL PROTECTED] writes:
I agree with you we should be fine deploying AAA records. I also agree
we need to be skeptical about A6 and if they continue to exist we need to
for sure do the type of experiments Randy suggested.
I must admit to being very skeptical of A6 -- my
I think it was recognized a long time ago that the initial deployment
of A6 records should be limited to two (or at most 3) levels. The question
is whether that is enough to avoid the horrors described by Dan Bernstein
over on IPNG.
'clever' people are likely to seriously abuse DNAME and
Hi Randy,
I stand corrected.
thanks
/jim
On Fri, 19 Jan 2001, Randy Bush wrote:
I agree with you we should be fine deploying AAA records. I also agree
we need to be skeptical about A6 and if they continue to exist we need to
for sure do the type of experiments Randy suggested.
A6 is PS not DS. If it will not work as is we have every right to oppose
it. I also agree with you about renumbering personally but I doubt even
if 10 of us joined forces and tried to alter the consciousness of the
collective IETF avalanche of renumbering belief systems which are valid
but may
DNAME for ISP.COM
DNAME for HOME.ISP.COM
DNAME for my.unit.HOME.ISP.COM
If ISP.COM renumbers all that has to be done is change the prefix of
the highorder bits of my Ipv6 address and only those SIG records
theoretically is the view (its much more complex), but this assumption
I think is bogus
but perhaps you should communicate TECHNICAL NEEDS to ngtrans
please explain what is non-technical about the need to have at least
one or two servers in place for ip6.arpa?
or does the lack of a specific number in Perry's request somehow
make it non-technical?
seems like the dnsops folks
Well, as I recall the earlier discussion, this rather depends on the expected
frequency and overhead of renumbering.
If the (frequency*overhead) of renumberings is comparable to the
(frequency*overhead) of lookups, we have to optimize for both.
Renumbering frequency depends on whether we
'clever' people are likely to seriously abuse DNAME and A6. we have already
seen unnecessary and confusing attempted use of DNAME over in the enum wg.
is there any *significant* advantage to them allowing more than one level of
indirection?
I can't tell that there is any significant
given that renumbering events are unlikely to be entirely automatic anyway
as fred succintly said, how often does renumbering occur? how often do
lookups occur? so, for which should we optimize?
makes sense to me.
Keith
y 18, 2001 4:48 AM
To: Bill Manning
Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED];
[EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: (ngtrans) Re: IPv6 dns
% an example of a worry is cache poisoning of an antique v4 bind.
%
% and there are thousands of vulnerable v4 b
"Christian Huitema" [EMAIL PROTECTED] writes:
Well, IPv6 users who want to access the IPv4 only DNS server will need
some form of solution. There is indeed no particular problem in
deploying a relay-resolver -- we are doing that all the time.
Ultimately, relay resolvers won't work. Our goal
By the way, the overall technical need is for the centralized
infrastructure of the net (which in practice turns out just to be our
DNS infrastructure) to properly support v6 users. That means we need
three things: proper support for reverse lookups, proper support for
v6 records in the NS
Randy Bush wrote:
I think it was recognized a long time ago that the initial deployment
of A6 records should be limited to two (or at most 3) levels. The question
is whether that is enough to avoid the horrors described by Dan Bernstein
over on IPNG.
'clever' people are likely to
At some point here we need to bite the bullet and put IPv6 DNS in
the ".". Why not an equal but separate set of root name servers
for V6 under IANA control? The named.root file can be distributed
in three flavors. IPv4 only, 6 and 4, 6 only? This is a political
issue and perhaps off the mark
Keith Moore [EMAIL PROTECTED] writes:
but perhaps you should communicate TECHNICAL NEEDS to ngtrans
please explain what is non-technical about the need to have at least
one or two servers in place for ip6.arpa?
Those would be nice, yes...
By the way, the overall technical need is for
% I think:
% - it should be okay to deploy records to certain degree
% - we need to be very conservative about deplying A6 records,
% as A6 presents highly different behavior than A/ records.
% # of queries would increase, # of additoinal records
During a trial, this can be done by operating on a trial subtree for
names. Something like example.ipv6dns.org. During the transition,
however, we want to progressively publish A6 records for the name
servers of regular domains. Obviously, any domain manager can do this on
their own initiative:
42 matches
Mail list logo
