[afnog] IPv6 - some interesting reading

[Phil Regnauld]

Geoff Huston (gih) writes:
> 
> 
>  not so - indeed I chaired one the early efforts back in 94 or so of an 
>  effort called "TACIT" that stood for transition and coexistence including 
>  testing. Then there was the NGTRANS working group which worked at this for 
>  many years. Transition took up about as much time as V6 itself. We did think 
>  hard about transition but....
>  transition was meant to happen well before IPv4 exhaustion
> 
>  oh well.

	[...]

>  no - it was industry leaving transition so late that all the assumptions 
>  about the transition are now flawed, but its too late to re-engineer the V6 
>  protocol, so we are in a little bit of a mess.

	Ok, thanks for the clarification.

> > 	Haven't read it yet, but I would encourage people to start
> > 	thinking about how they'll actually deploy *and* migrate
> > 	from v4 to v6.
> 
>  No - its worse. The transition is DUAL stack - for the next (n) years you 
>  need to have BOTH v6 and V4. Its NOT an either or. You need BOTH for as long 
>  as this dual stack transition is underway.

	Right.

>  Quitting out of V4 only happens when we've managed to get _all_ the V4 
>  legacy to dual stack.

	Indeed, since all new v6-only installations will still need a way
	to talk to v4-only sites, right ?

>  So we are _all_ going to need V4 for new deployments as well as existing 
>  deployments for much much longer than 2 years if we want One Internet to 
>  continue to work as One Internet.