On 31 Dec 2008, at 20:15, William Herrin wrote:
I would like to hear about additional NAT-based techniques which, like Chris' notion, suggest a credible plan for reducing the demand for core routing slots compared to the status quo.
I cited "NAT66" specifically. I also noted both where that discussion has been happening (although some related discussion has also happened here) and the existence of an I-D. For example, one form of NAT66 is stateless, meaning that sessions could be initiated from outside or inside the NAT perimeter. Please go read the draft. I also specifically cited "architected NAT", where there is a concrete specification that has particular properties. Thirdly, I've talked in the past about how Locator rewriting (which is a kind of NAT) could be used with ILNP. This is a bit similar to Six/One, but is not identical. This is discussed in detail in various published papers. Either of these is rather different from your apparent view of a single monolithic "IPv4 NAT". In fact, a wide range of different forms of NAT exist, most of which are not very well documented. Different NAT techniques/implementations have different properties.
I would prefer not to hear about the usefulness of vanilla NAT.
Certainly, it is your choice what you choose to read or hear.
We have determined beyond the shadow of a doubt that the status quo is not, at this time, self-reducing.
The RG has *not* made a determination that all existing NAT techniques (there are a wide range of various NAT methods available in myriad different implementations) aren't part of the possible solution set. Under IRTF process rules, only the RG Chairs have the authority to make a consensus determination.
Vanilla NAT is indeed useful, but as widely deployed today it is already a part of the status quo.
Since IPv4 NAT has so many different forms, with varying properties, I have no idea what you mean when you say "Vanilla NAT". Yours, Ran [email protected] _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
