Re: [BEHAVE] [v4tov6transition] Any Experience with Using Behave's StatelessNAT-PT for IMS-SIP VoIP Application...

[Brian E Carpenter <brian.e.carpenter@gmail.com>]

<AANLkTin1gQhnS=w4+ehLyJOyO+mz0NNnz-4rD-N97ZEq@mail.gmail.com>
<EC91E98C3BC6A34B917F828067B9335C1535E733DE@PRVPEXVS07.corp.twcable.com>	<005301cb5a6f$75490320$5fdb0960$@com>
<AANLkTi=+Fx4c6UvcBi3GO2U2Cg9y7EkkTcmyO2q_YQ6t@mail.gmail.com>
<576604.55924.qm@web111407.mail.gq1.yahoo.com>
<179c01cb5ac3$db7a77b0$926f6710$@com>
<0F1D6CA3-F978-4984-90B3-CE6701BDB457@cisco.com>
In-Reply-To: <0F1D6CA3-F978-4984-90B3-CE6701BDB457@cisco.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Sender: owner-v6ops@ops.ietf.org
Precedence: bulk
List-ID: <v6ops.ops.ietf.org>

On 2010-09-24 07:19, Fred Baker wrote:
> On Sep 22, 2010, at 7:05 PM, Dan Wing wrote:
> 
>>> Cameron, NAT64 also requires ALGs.
>> No, only for protocols that don't do their own NAT traversal. 
> 
> If we agree that some don't, then you are agreeing with each other. Protocols that do their own NAT traversal (anything that keeps application layer stuff at the application layer and network layer stuff at the network layer) are fine. Protocols that mix the two - FTP, SIP, perhaps SMTP, etc - require some form of ALG or proxy to fix up the application layer stuff. Simplest approach is an application that speaks IPv6 on one side and IPv4 on the other and terminates the sessions, such as an SMTP MTA; if relays don't exist, it gets messy, but then the NAT world has always been messy.

Shameless plug: there's another way to look at this topic:
draft-carpenter-referral-ps-01.txt

   Brian