Re: draft-chen-v6ops-nat64-experience-02

[Behcet Sarikaya <[e-mail]>]

On Mon, Jul 9, 2012 at 12:38 PM, Cameron Byrne <[e-mail]> wrote:
> On Mon, Jul 9, 2012 at 10:18 AM, Behcet Sarikaya <[e-mail]> wrote:
>> On Fri, Jul 6, 2012 at 9:59 AM, Michael Richardson
>> <[e-mail]> wrote:
>>>
>>>>>>>> "Aleksi" == Aleksi Suhonen <[e-mail]> writes:
>>>     Aleksi> Within an hour, all the IPv4 addresses in the pool for our
>>>     Aleksi> NAT64 were registered to this one device.
>>>
>>> Do I understand that you attempt to provide a single IPv4 address 1:1
>>> with a an internal IPv6 address? (NAT vs NAPT)
>>
>> It seems like this is what is called stateless NAT64.
>> I am not sure if there is any document specifying stateless NAT64?
>>
>> Regards,
>>
>> Behcet
>
> Stateless = http://tools.ietf.org/html/rfc6145

Are you sure?

Here is a quote from 6146:
Stateful NAT64 is a mechanism for translating IPv6 packets to IPv4
   packets and vice versa.  The translation is done by translating the
   packet headers according to the IP/ICMP Translation Algorithm defined
   in [RFC6145].

Regards,

Behcet

>
> Stateful = http://tools.ietf.org/html/rfc6146
>
> If the goal is providing a dynamic access from an IPv6-only network
> toward IPv4-only internet, RFC 6146 is the optimal choice.
>
> RFC 6145 has limited use for the cases for IPv6 - > IPv4 since it is
> 1:1 mapping.  Most people do IPv6 because IPv4 is limited, so ...
> doing 1:1 mapping does not really buy you anything.  You can just use
> IPv4 and achieve the same scale.
>
> The best use case i have seen for RFC 6145 is for the data center
> environment http://fud.no/talks/20120417-RIPE64-The_Case_for_IPv6_Only_Data_Centres.pdf
>  as well as the mapping of the entire IPv4 internet into IPv6 as is
> the case of 464XLAT CLAT in the IPv4->IPv6  scenario.
>
> CB