Re: [v6ops] Please review the No IPv4 draft

[Doug Barton <[e-mail]>]

On 04/14/2014 12:42 PM, Ted Lemon wrote:
> On Apr 14, 2014, at 2:30 PM, Owen DeLong <[e-mail]> wrote:
>> Seems to me that the same potential for abuse exists with either situation.
>
> This is true, and the remedy exists as well.  It's worth pointing out that "please review" doesn't mean "please redesign this."   It means "please point out technical flaws that you see."   None of what you or Nick has said sound like technical flaws to me—they sound like "I would prefer to do it this other way."

Ted,

Sometimes ideas that are cooked up in one environment don't stand the 
light of day when they escape the environment they were created in. I 
think this is one of those times.

First, any draft that suggests adding new options to RA is going to get 
a negative reaction from me, for reasons I've stated ad nauseum in the 
past, and I really don't want to open that can of worms on this point.

Second, as others have pointed out this is layering violation of 
monumental proportions, and thinking about this as an implementor (even 
if the RA stuff was dropped) I would not be keen to try and navigate 
this. Services which start up listening on IPv4 networks which suddenly 
have those networks yanked out from under them comes immediately to 
mind, but I haven't even started thinking hard about the problem(s) yet.

Third, this draft doesn't pass a simple laugh test. If you don't want 
IPv4 on your network, configure it away ... and you're done. There is no 
need for any of this.

Fourth, it doesn't pass the "stop tinkering with the protocol and let 
the installed base settle on a feature set" test.

Finally, and please pardon me for being brutally honest, this whole 
concept seems childish. When I saw the subject line I was deeply 
concerned that this was going to turn out to be a "We hate IPv4 so much 
we want a knob to make it go away" type of thing. As much as the authors 
may not have intended it that way, that's exactly how it reads.

So with all due respect to the authors, and the working group, this is a 
terrible idea, and the draft should be allowed to expire.

Doug