Heiner,
I do not know of any point where the problem you describe was
stated, or intended to be addressed by ISP or most o the other proposals
that the RRG looked at. It is not within the scope of the LISP WG.
Yours,
Joel
On 1/19/2013 5:50 AM, [email protected] wrote:
Joel,
I honestly thought that LISP would maintain/provide global uniqueness by
the doublet
{LISP-site unique IPv4 address of the EID; Locator} and only in case my
memory doesn't deceive me it was once (long ago) a pro-argument of LISP.
Given this assumption you must admit that only LISP-DNS would work, not
-ALT, nor -DDT, and that under such a presumption my emails would have
really mattered for THIS group.
By Dino's yesterday email I learned that LISP-DDT presumes global
uniqueness of the IPv4 address which means
more NAT, and not less NAT. He also mentioned IPv6-NAT which can pretty
easyily be predicted: Company XYZ switches from IPv4 to IPv6 but wants
to keep its NAT.
The result: Complexity induces even more complexity.
Facit: At least myself I have learned that NAT is the ultimate solution
( and not LISP-DDT) for this particular problem.
Heiner
-----Ursprüngliche Mitteilung-----
Von: Joel M. Halpern <[email protected]>
An: heinerhummel <[email protected]>
Cc: lisp <[email protected]>
Verschickt: Fr, 18 Jan 2013 11:51 pm
Betreff: Re: [lisp] LISP and IPv4
Heiner, I am having trouble determining what you would like the working
group to do with your input.
You keep mentioning TARA. There is no Internet-Draft for TARA, and even
if there were, it would be out of scope for this working group.
Your concerns about overlapping addresses, seem to be aimed at use cases
which while interesting are not within the charter of the working group.
Dino has nonetheless gone out of his way to address your comments.
What are you looking for?
Yours,
Joel M. Halpern, speaking as co-chair
On 1/18/2013 5:46 PM,[email protected] <mailto:[email protected]> wrote:
Dino,
In spite of adding 4 locator octets you still need NAT !:-( C'mmon!
And you try to sell it as a great feature! :-(
Your position is: Thanks to NAT you have no ipv4 address depletion
problem, right ?!
NAT is indeed a great bail out (like GRE which I mentioned lately), but
nevertheless a poor and cumbersome solution.
And there are many IETF goers who know much much better than me about
NAT's uglyness. And combined with LISP's servers infrastructure
solutions become even more cumbersome.
So I admit: NAT will keep LISP alive. But also: I expected LISP to be
better: Having an IPv4 address plus a locator there is no need to depend
on NAT i.e. on additional TCP octets in order to provide global uniqueness.
LISP will cement the bad IPv4 AND IPv6 paradigms. Dino, you shouldn't
point to IPv6 and say that IPv6 shall/will do better.
In the past IPv6 has always emulated whichever IPv4 protocol. And IPv6
will rather emulate LISP than come up with better routing.
You also wrote:
"We do not want to change DNS. We do not require any new support from DNS."
Well, TARA as outlined by draft-tara-hummel-00.txt only required DNS
service as by RFC1712.
Yes, Its current status however requires some extra DNS service which is
to provide also a host's "home geopatch number"
This I need for stateless multicast whereby the mobile sender is roaming.
Do you think there will ever be such models with LISP ? So far you do
not have a Multicast-Locator nor an Anycast Locator!
Regards
Heiner
_______________________________________________
lisp mailing list
[email protected] <mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/lisp
_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp
