[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: WG last call on NAT-PT to Experimental
----- Original Message -----
...
From: "Baker Fred" <fred@cisco.com>
To: "Rémi Després" <remi.despres@rd-iptech.com>
Sent: Monday, March 14, 2005 1:20 PM
Subject: Re: WG last call on NAT-PT to Experimental
...
I'm not opposed to necessary work being done, but I'm not so sure what work
needs to *be* done. We seem to have a lot of transition efforts around, but
there must be a way to bring them together into a single simple solution.
Bring us that one simple path, and then we'll talk.
What I am discussing can become a piece of any "simple solution" which
integrates ISATAP.
The issue can be restated as follows:
GOAL? : is it desirable that IPv4-only devices, at IPv4 fixed
addresses in dual-stack sites, can be called in a standard way, at least for
applications which are compatible with IPv4-NATs, by ordinary IPv6-capable
hosts which consult the DNS?
SOLUTION? : Is there a solution for this GOAL, if accepted, which is
compatible with all currently recommended IETF specifications, and simple
to implement and to operate ?
IMHO the answer is YES to both questions.
Discussing in IETF details of the SOLUTION, while no consensus is
available on (1) would be premature.
The FIRST QUESTION to be answered should then be:
"Is the above GOAL found generic enough by enough people to deserve
IETF attention? And if yes, where should it be discusssed?"
Yet, just for information at this stage (please skip if there is not any
interest for the GOAL), here is, after having studied a few other schemes,
my preferred SOLUTION, based on a simple extension of ISATAP:
- The fact that a called device is IPv4-only, as opposed to an
ordinary ISATAP host, is recognizable at site entrance by a Host Protocol
Indicator (HPI) in the called IPv6 address itself.
- The HPI is proposed to be in a bit which is always a 0 in current
ISATAP addresses (set to 1 for IPv4-only devices). The proposed bit is
the low order one in the first octet of the ISATAP IPv4-address field. (This
bit is 0 in all IPv4 private prefixes defined in RFC 1918. Thus, 10/8,
172.16/12 and 192.168/16 prefixes, used for IPv6-capable hosts, become
11/8, 172.17/12 and 192.169/16 for IPv4-only ones).
- For IPv4-only called devices, a "NAT6to4" function performs the
required translation between IPv6, outside, and IPv4, inside.
- This NAT6to4 function is a quite simplified version of the generic
NAT-PT: No ALG, in particular no DNS ALG; no fragmentation support; not
even port translation assuming that enough local IPv4 addresses, being in
the private space, can be allocated to NAT6to4. (With these simplifications,
NAT6to4 is not subject to operational problems of NAT-PT identified in
draft-ietf-v6ops-natpt-to-exprmntl-00.txt.)
Regards,
Rémi