Noel Chiappa wrote:
From: Keith Moore [EMAIL PROTECTED]
what we really need is a layer of indirection at the BGP level so that
sites can have stable addresses without having to NAT.
You mean, have a namespace for use by the path-selection algorithms, one
which is separate
what we really need is a layer of indirection at the BGP level so that
sites can have stable addresses without having to NAT.
we should rather drop stable address requirement by having session
layer protocol (something better than TCP).
having a session layer protocol
On Fri Aug 24 01:48:00 2007, David Conrad wrote:
I'll take ease in renumbering over application transparency for
any
large network.
I find this confusing as a concern - how often do you renumber?
How often do you want to change service providers?
Well, I have a pretty good one, so not
, August 23, 2007 9:10 PM
To: Stephen Kent
Cc: Hallam-Baker, Phillip; RJ Atkinson; Sam Hartman; ietf@ietf.org
Subject: Re: The Internet 2.0 box Was: IPv6 addresses really
are scarce after all
The DNS is a 1980's technology. We used hosts.txt prior to that.
yeah, that was a typo. (and I do
Keith Moore [EMAIL PROTECTED] writes:
DNS is the Achilles heel of the Internet. it's way too unreliable, too
hard to configure correctly, too often out-of-sync with the real world.
it's not extensible enough.
DNS is surely the worst global naming system ever invented, except for
all the
From: Keith Moore [EMAIL PROTECTED]
I think it can be done without changes to IPv6, since it doesn't affect
the packet format, and the only things that have to know about it are
routers and network management tools.
I'm not sure I follow you. Are you talking about what we've
I try to learn from past efforts - both negative and positive. You on the
other hand demand that we consider the 1983 design of the Internet as
sacrosanct, except of course when you are sneering at people for proposing
'1980s technology'.
Okay, fair enough. Actually the Internet
Railing against the shortcomings of the current DNS (or any current
technology, for that matter) does little to get us to a better system.
If you know of a better approach, what are you doing to make it a
reality?
The purpose of my argument was to dispel the notion that DNS should be
Keith Moore wrote:
[..]
I believe I understand how to replace DNS with a better protocol while
preserving the existing hierarchy and RRsets and DNSSEC, and allowing
graceful transition from the old to the new. However, I'm not sure that
I have enough understanding of DNS's failings to
Bickering about all this is fun
of course, but it doesn't help coming to a solution, especially as the
solution doesn't have a defined problem set and what it is supposed to
solve.
of course. but the purpose was not to bicker, but rather to do some
damage control - to try to discourage
Dave,
On Aug 24, 2007, at 1:32 AM, Dave Cridland wrote:
I'm honestly struggling to see what the issue is here. I certainly
agree that renumering is a pain, but I don't follow why renumbering
is so significantly painful that it's worth breaking the network
for. I'm not saying it isn't, I
On 24-aug-2007, at 17:28, David Conrad wrote:
If you obtain address space from a service provider and you decide
to change providers, you have (in most cases) two options: renumber
or deploy NAT.
Nonsense.
Assuming you're not going to take the address space with you (which
is not a
The IETF has a simple process for all of this: write a draft.
Not true.
The IETF also runs a large number of mailing lists for discussion of
things both general and specific. It is not necessary to start work by
writing a draft. One can also start work by discussing the problem area
on one or
On 8/24/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
snip
No reason to attack him like you did and I specifically want to address
this because mailing lists have a much larger audience than their
participants. If such attacks are not answered it creates barriers for
new blood to enter into
On Aug 24, 2007, at 8:46 AM, Iljitsch van Beijnum wrote:
On 24-aug-2007, at 17:28, David Conrad wrote:
If you obtain address space from a service provider and you decide
to change providers, you have (in most cases) two options:
renumber or deploy NAT.
Nonsense.
Sigh. I forgot to be
Roger Jørgensen wrote:
On 8/24/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
snip
No reason to attack him like you did and I specifically want to address
this because mailing lists have a much larger audience than their
participants. If such attacks are not answered it creates barriers for
On 8/24/07, David Conrad [EMAIL PROTECTED] wrote:
snip
If you obtain address space from a service provider and you decide to
change providers, you have (in most cases) two options: renumber or
deploy NAT. It is a simple cost/benefit tradeoff, with the costs
impacting software and protocol
On 24-aug-2007, at 18:44, David Conrad wrote:
If you obtain address space from a service provider and you
decide to change providers, you have (in most cases) two options:
renumber or deploy NAT.
Nonsense.
Sigh. I forgot to be pedantic and use the IETF-mandated terminology.
If you
Iljitsch,
On Aug 24, 2007, at 10:03 AM, Iljitsch van Beijnum wrote:
Regardless of the theatrics, this statement is still incorrect. As
I said in my previous message, you can't keep the old addresses
internally either so all of this buys you nothing.
I suspect the number of people who NAT
There seems to be consensus that trying to stop NAT in the v4 world is
futile. Good. So then we ask: what will keep it from happening in
the v6 world?
I postulate the following as one necessary, and perhaps sufficient,
condition:
IN ORDER TO AVOID v6 NAT: Network administrators of any home
nice idea, but I'm fairly convinced that it's impractical. there are
just too many interfaces, many of them nonstandard and application
specific, that need to know about IP addresses.
maybe we could come up with a 90% solution, but that 10% is still a bear.
I'm back to thinking that we have to
IN ORDER TO AVOID v6 NAT: Network administrators of any home or
enterprise network need to have, at essentially zero cost, ownership
or control over SOME NUMBER of bits of the v6 address space,
sufficient to uniquely address each host in their network, and such
that a change in ISP or
from realworld experience in providing IPv6 services at an ISP,
and as a customer of that service:
/48 PA assignments to the customer is sufficient.
for roaming clients (like travelling laptops with PPP) there's a
different requirement.
IMHO even a
/48 PA assignments to the customer is sufficient.
for roaming clients (like travelling laptops with PPP) there's a
different requirement.
IMHO even a traveling laptop with PPP needs to be able to subnet.
I've lost count of the number of times I've needed to do this in
IMHO even a traveling laptop with PPP needs to be able to subnet.
I've lost count of the number of times I've needed to do this in IPv4
but been stuck with a single /32. I have also lost count of the number
of times I've used a laptop as a router in the days when I had a /28
routed to
On 8/24/07, Keith Moore [EMAIL PROTECTED] wrote:
nice idea, but I'm fairly convinced that it's impractical. there are
just too many interfaces, many of them nonstandard and application
specific, that need to know about IP addresses.
maybe we could come up with a 90% solution, but that 10% is
that Kieth is no longer opposing IPv4 NAT which is something.
-Original Message-
From: Keith Moore [mailto:[EMAIL PROTECTED]
Sent: Tuesday, August 21, 2007 12:46 PM
To: Hallam-Baker, Phillip
Cc: Sam Hartman; RJ Atkinson; ietf@ietf.org
Subject: Re: The Internet 2.0 box Was: IPv6 addresses
Hallam-Baker, Phillip wrote:
If we can meet the needs of 80% of Internet users with some form of shared
access there will be more addresses left for the 20% with greater needs.
with 2**128 potential addresses, this is not only unnecessary, it's
harmful. there's far greater benefit to be
Keith == Keith Moore [EMAIL PROTECTED] writes:
Keith Hallam-Baker, Phillip wrote:
If we can meet the needs of 80% of Internet users with some
form of shared access there will be more addresses left for the
20% with greater needs.
Keith with 2**128 potential
On Thu, 23 Aug 2007, Keith Moore wrote:
basically DNS is not the sort of
thing you want to saddle every application in the Internet with,
[...]
It would seem that you are 20+ years too late. Just what color are
the Frogstar fighters in your universe, anyway? [1]
Matt
[1]
At 11:23 AM -0700 8/23/07, Hallam-Baker, Phillip wrote:
If we can meet the needs of 80% of Internet users with some form of
shared access there will be more addresses left for the 20% with
greater needs.
I suspect that the actual percentages are more like 95% and 5%.
My Internet use is
On Thu Aug 23 21:12:17 2007, Sam Hartman wrote:
Keith == Keith Moore [EMAIL PROTECTED] writes:
Keith Hallam-Baker, Phillip wrote:
If we can meet the needs of 80% of Internet users with some
form of shared access there will be more addresses left for
the
20% with greater
I'll take ease in renumbering over application transparency for any
large network.
I find this confusing as a concern - how often do you renumber?
How often do you want to change service providers?
Regards,
-drc
___
Ietf mailing list
Ietf@ietf.org
Sam Hartman wrote:
Keith == Keith Moore [EMAIL PROTECTED] writes:
Keith Hallam-Baker, Phillip wrote:
If we can meet the needs of 80% of Internet users with some
form of shared access there will be more addresses left for the
20% with greater needs.
Larson, Matt wrote:
On Thu, 23 Aug 2007, Keith Moore wrote:
basically DNS is not the sort of
thing you want to saddle every application in the Internet with,
[...]
It would seem that you are 20+ years too late. Just what color are
the Frogstar fighters in your universe, anyway?
The DNS is a 1980's technology. We used hosts.txt prior to that.
yeah, that was a typo. (and I do remember using hosts.txt)
though somehow, 1980s technology doesn't sound a lot better.
___
Ietf mailing list
Ietf@ietf.org
From: Keith Moore [EMAIL PROTECTED]
what we really need is a layer of indirection at the BGP level so that
sites can have stable addresses without having to NAT.
You mean, have a namespace for use by the path-selection algorithms, one
which is separate from the namespace used for
what we really need is a layer of indirection at the BGP level so that
sites can have stable addresses without having to NAT.
we should rather drop stable address requirement by having session
layer protocol (something better than TCP).
itojun
Why is Keith so desperately wedged on one particular means of achieving his
objective?
It is entirely possible to make peer to peer applications work well with NAT,
it is entirely possible even to make a server application work well with NAT.
We are running out of IPv4 addresses and it is
Hallam-Baker, Phillip wrote:
Why is Keith so desperately wedged on one particular means of achieving his
objective?
because it's by far the simplest and most reliable means available.
It is entirely possible to make peer to peer applications work well with NAT,
it is entirely possible
On 21-aug-2007, at 18:39, Hallam-Baker, Phillip wrote:
It is entirely possible to make peer to peer applications work well
with NAT, it is entirely possible even to make a server application
work well with NAT.
It is entirely possible to make it so that you can breathe under
water. When
41 matches
Mail list logo