Re: [v4v6interim] [46translation] [BEHAVE] Proposal for new BEHAVEcharter

On 23 okt 2008, at 18:58, Alain Durand wrote:

> IPv6 is just like IPv4 with longer addresses. This is where deployment
> reality trumps architectural purity.

Disagree. Two reasons:

1. The longer addresses don't just make a quantitative difference, but  
also a qualitative difference. Things like SEND, HIP and shim6 aren't  
possible with IPv4 because they put crypto stuff in spare address  
bits. Things like port overloading NAT aren't necessary in IPv6  
because there is abundant address space.

2. If we don't use the opportunity that IPv6 gives us to fix some of  
the brokennes of the current IPv4 internet, we're not going to get  
another chance any time soon. If we use IPv6 just like IPv4 we might  
as well stay with IPv4. We know that with enough NAT it actually is  
possible to squeeze blood from stones.

> What I've seen in the last 3 years is that it is much, much easier  
> to deploy
> IPv6 if you keep the exact same operational model as in IPv4.

But then we've gained nothing.

> Many folks are considering NAT44 as a simple solution where they  
> just have
> to put a box in front of their network to make whatever problem they  
> think
> they have go away. I can easily see that the same people would like  
> to do
> the same thing with IPv6. Which exact problem they were solving is
> irrelevant to this discussion, the point is about preserving the same
> operational model.

Let me try a more opportunistic argument on you: in IPv4, you can't be  
pedantic because then you're disconnected. With IPv6, you can not  
route, filter, break whatever you don't like and the offending party  
then falls back on IPv4 so you still have connectivity. For this  
reason, it's very likely that there will be a lot more operational  
hard ball with (perceived) architectural purity with IPv6. This means  
that the perfect storm of different kinds of users that all saw NAT as  
a solution to their problem and thus required application makers to  
work around NAT won't happen with IPv6, and the onus of implementing  
the workarounds will fall solely on the party implementing the NAT for  
a very long time to come. This is of course often impossible with a  
port overloading NAT so a lot of things that work with IPv4+NAT will  
fail with IPv6+NAT.
_______________________________________________
v4v6interim mailing list
v4v6interim@ietf.org
https://www.ietf.org/mailman/listinfo/v4v6interim