On Fri, 23 Jul 2010, Marco Hogewoning wrote:
In short, why a /48 'Because we can!'.
I do not buy your argument "consumers expect a /48 so we'll get grief if
we don't give it to them." As others have pointed out, "consumers" don't
want IPv6, they want web surfing, playing games, and e-mail.
When this topic has come up in the past I've posted my discomfort on the
idea of "/48s all the way down" and given that there is now good
traction for the idea of actually using the squishy stuff between our
ears for designing address plans, I'm not going to rehash that.
What I will rehash, because no one else has repeated it yet, is that
there is a middle ground between "assign /{56|60} only, and then suffer
later if it's not enough" and "/48 all the way down!" You simply RESERVE
a /48 for each customer, but ASSIGN the first /56 in the range. This
will give you lots of great operational experience in how your customers
actually use IPv6 under the umbrella of "/56 really is likely to be
enough" while not having painted yourself into any corners if it turns
out we're wrong about that. Why those particular boundaries? There are
256 /64 subnets in a /56 (once again, should be enough), and 256 /56s in
a /48 (a: I like round numbers, b: see below).
Once you've had some operational experience with this plan it should be
pretty obvious how to move forward. If it turns out that /56 really is
enough and you're running out of /48s to reserve, you go back through
and start again on the /49 boundary of the /48s you already reserved. If
it turns out that your customers really do need /48s, you go back to the
RIR and ask for more space.
Personally, I think the right answer is going to be a mix of both, but
the beauty of this plan is that it allows for that.
hth,
Doug
--
Improve the effectiveness of your Internet presence with
a domain name makeover! http://SupersetSolutions.com/
Computers are useless. They can only give you answers.
-- Pablo Picasso
