Re: draft-bourbaki-6man-classless-ipv6-00

[Roger Jørgensen <[e-mail]>]

sorry for the language in this email, and it's not direct at you
Brian, it's toward us all.

On Mon, Jun 5, 2017 at 1:05 AM, Brian E Carpenter
<[e-mail]> wrote:
<snip>
> ......... The point is to establish
> that routing is classless and /64 is a parameter of specific
> addressing schemes.
<snip>

If we here in this group forget that most people that will use IPv6
don't care at all
about anything in any RFC, they actual don't really care that there are a RFC
system - what they care about is that their daily life is easy. This
_I_ think is very
true for the enterprise world. They go for quick and dirty, and
cheapest solution.

So, for most of them they will only hear and see "IPv6 is classless",
and with their
IPv4 background they _will_ start todo all kind of crazy stuff, all
the things WE
seems to think we can handle by putting in extra words in a RFC.
Really - if we are stupid, yes that's the word, that we think any
words we put into
an RFC will prevent NAT66, /128 and /120 and all the crazy things people do in
IPv4 and now will be doing in IPv6, then well ... let's just rename
IPv6 into IPv4+
and be done with it.

So sure we will get more IPv6 deployment, but NOT the kind that actual
will solve
application layer problem created by NAT & alike, and we are not any
better of at
all. Not to forget that we very likely are to trigger a routing/policy
change because
there are no hard /64 anymore......


This one universal /64 is what almost every none-technical, and most technical
I've ever meet understand about IPv6. Some think it's a waste of space but it's
accepted and they work with it.



With that said - sure I agree with Brian on that one sentence.

I actual also agree with the intent of the draft, no hardcoded /64's anywhere
(from my lenghty discussion with Job online over the last few days), it should
be DEFAULT /64, but changable.
I just can't accept the followups that "classless & IPv6" will create since most
people would never read the finer details of what that actual mean and
we will have IPv4+, not IPv6.
But if most people are fine with the draft and don't belive it will cause the
havoc I think it will, then count me in the rough, I'm fine with that.


Also, I think I remember that most hardware are optimized around that
/64 boundary, how will this affect newer chipset and hardware? Should they
optimize their hardware to route anything from /8 to /128? Sounds very
expensive and not something that will make IPv6 more popular.



-- 

Roger Jorgensen
[e-mail] / [e-mail]