On 6-okt-2007, at 7:00, Michel Py wrote:
Think about the following: even if in 3 years 50% of hosts were
IPv6-only capable, it would not diminish the need for IPv4. All the
double-NAT tricks, unused address reclaim, config cleanup etc are
to happen now no matter what. I'm not saying it's going to be easy or
cheap, but as long as there is a need for v4 it will happen.
The unanswered question is: are all these tricks going to be enough to
keep operating IPv4. Nobody knows, but almost everyone who already
v4 address can wait.
Well, if in the forseeable future (3 years is a bit short, though)
50% of all hosts has IPv6 connectivity, I would call that a
resounding success. (I'll even take 25 or even 10 % or whatever is
enough to make most ISPs deploy IPv6 in their networks.) That the
other 50/75/90% is still IPv6-only wouldn't be a problem: presumably,
IPv4 works for them so there is no need to add IPv6.
The tricky part is what happens to people that run into limitations
that exist in IPv4 but not in IPv6. (NAT, hard to get enough
addresses, that kind of stuff.) So far, deploying IPv6 to work around
these problems has rarely been a workable option. But hopefully, it
will become one in the next few years.
Ietf mailing list