ietf
[Top] [All Lists]

RE: How the IPnG effort was started

2004-11-23 18:40:53
At 12:07 PM 11/21/2004, Peter Ford wrote:
Content-class: urn:content-classes:message
Content-Type: multipart/alternative;
        boundary="----_=_NextPart_001_01C4CFEC.A4503CD1"


Noel,

You are sorely under-representing the IETF's and your own efforts wrt NATs. I think of your taxonomic study of NATs much in the same vein as Carl Linnaeus's "Systema Naturae".

In fact, given the intellectual contributions by the think tank inside the IETF to NATs, the working groups on NATs, and the protocol engineering for NATs devised in the IETF, one would think that we can credit the IETF with NATs and the emergent Internet NAT architecture, guided by the IESG and overseen by the IAB.

The IETF NAT Working Group was formed to document the many technologies which were being labelled as NAT and provide some common reference. We also produced documentation on how to design protocols that can transit NATs without help, and document what things were broken by NAT. The WG did NOT develop NAT.

The IETF did, however, have a lot to do with limiting the availability of IPv4 address space, which pushed people to consider building NAT devices. Proteon did a lot of work in 1993 - 1994.


One of the more interesting things we may encounter in the post-IPv4 era is a great simplication, but not elimination of NATs. In other words: NATs become strictly address re-writers for IPv6 addresses. And yes, some of the ongoing research in NAT architecture will probably make it into the IPv6 world. I can easily imagine a world where hosts use NATs and IPv6 simultaneously and I suspect this might be a next-gen firewall technology. People seem to forget that people buy NATs for IP address sharing and firewalling. They don't seem to "get it" that there are very few people who would ever buy a NAT because of IPv4 address limitations.

Actually, I expect NATs will survive for a very long time as the low-cost way to multihome. Many devices are on the market today which provide multiported NAT with 2 or more upstream ISPs routing a single address or subnet to a port on the device, and with the device performing NAT, load balancing and other functions all in one box. Certainly this will be a simpler solution for users within many organizations than trying to deploy the multi6 schemes with multiple addresses on every machine.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf
<Prev in Thread] Current Thread [Next in Thread>